├── requirements.txt ├── _config.yml ├── assets ├── css │ └── styles.scss ├── favicon.png ├── preview.png └── js │ └── script.js ├── .gitattributes ├── Gemfile ├── start.bat ├── .gitignore ├── run-repomix.bat ├── _layouts └── default.html ├── _includes └── footer.html ├── LICENSE ├── _data ├── links.yml ├── timeline.md └── timeline.yml ├── feed.xml ├── Gemfile.lock ├── .github └── workflows │ └── deploy.yml ├── _sass ├── _variables.scss └── _layout.scss ├── index.md ├── README.md ├── notes.md └── scripts └── convert_timeline_events.py /requirements.txt: -------------------------------------------------------------------------------- 1 | PyYAML -------------------------------------------------------------------------------- /_config.yml: -------------------------------------------------------------------------------- 1 | url: "https://nhlocal.github.io" 2 | baseurl: "/AiTimeline" -------------------------------------------------------------------------------- /assets/css/styles.scss: -------------------------------------------------------------------------------- 1 | --- 2 | --- 3 | @use "variables"; 4 | @use "layout"; 5 | -------------------------------------------------------------------------------- /.gitattributes: -------------------------------------------------------------------------------- 1 | # Auto detect text files and perform LF normalization 2 | * text=auto 3 | -------------------------------------------------------------------------------- /assets/favicon.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/NHLOCAL/AiTimeline/HEAD/assets/favicon.png -------------------------------------------------------------------------------- /assets/preview.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/NHLOCAL/AiTimeline/HEAD/assets/preview.png -------------------------------------------------------------------------------- /Gemfile: -------------------------------------------------------------------------------- 1 | source "https://rubygems.org" 2 | gem "jekyll" 3 | gem "csv" 4 | gem "logger" 5 | gem "base64" -------------------------------------------------------------------------------- /start.bat: -------------------------------------------------------------------------------- 1 | python scripts/convert_timeline_events.py _data/timeline.md 2 | 3 | bundle exec jekyll serve -------------------------------------------------------------------------------- /.gitignore: -------------------------------------------------------------------------------- 1 | _site 2 | .sass-cache 3 | .jekyll-cache 4 | .jekyll-metadata 5 | vendor 6 | repomix-output.xml 7 | repomix-output.md 8 | -------------------------------------------------------------------------------- /run-repomix.bat: -------------------------------------------------------------------------------- 1 | @echo off 2 | 3 | npx repomix --style markdown --remove-comments -i "scripts/**,notes.md,**/timeline.md" 4 | 5 | pause -------------------------------------------------------------------------------- /_layouts/default.html: -------------------------------------------------------------------------------- 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | {{ page.title }} 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | {{ content }} 19 | 20 | 21 | 22 | -------------------------------------------------------------------------------- /_includes/footer.html: -------------------------------------------------------------------------------- 1 |
2 |
3 |

About

4 |

This open-source timeline tracks the rapid evolution of Artificial Intelligence. Maintained by NHLOCAL and the community.

5 | 6 | Contribute 7 | 8 |
9 | 10 |
11 |

Project

12 | 23 |
24 |
-------------------------------------------------------------------------------- /LICENSE: -------------------------------------------------------------------------------- 1 | MIT License 2 | 3 | Copyright (c) 2024 NHLOCAL 4 | 5 | Permission is hereby granted, free of charge, to any person obtaining a copy 6 | of this software and associated documentation files (the "Software"), to deal 7 | in the Software without restriction, including without limitation the rights 8 | to use, copy, modify, merge, publish, distribute, sublicense, and/or sell 9 | copies of the Software, and to permit persons to whom the Software is 10 | furnished to do so, subject to the following conditions: 11 | 12 | The above copyright notice and this permission notice shall be included in all 13 | copies or substantial portions of the Software. 14 | 15 | THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR 16 | IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, 17 | FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE 18 | AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER 19 | LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, 20 | OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE 21 | SOFTWARE. 22 | -------------------------------------------------------------------------------- /_data/links.yml: -------------------------------------------------------------------------------- 1 | - text: The Latest AI Innovations 2 | url: https://www.futurepedia.io/ai-innovations 3 | - text: Artificial Intelligence Index 4 | url: https://aiindex.stanford.edu 5 | - text: Analyzing 2023's Milestones and Forecasting 2024's Trends 6 | url: https://masterofcode.com/blog/ai-highlights-2024 7 | - text: Timeline of Artificial Intelligence - Wikipedia 8 | url: https://en.wikipedia.org/wiki/Timeline_of_artificial_intelligence 9 | - text: The History of Artificial Intelligence - Complete AI Timeline 10 | url: https://www.techtarget.com/searchenterpriseai/tip/The-history-of-artificial-intelligence-Complete-AI-timeline 11 | - text: 'From the World Wide Web to AI: 11 Technology Milestones That Changed Our 12 | Lives' 13 | url: https://www.weforum.org/agenda/2024/03/11-technology-milestones-ai-quantum-computing-vr/ 14 | - text: 'Artificial Intelligence (AI) and ChatGPT: History and Timelines' 15 | url: https://www.officetimeline.com/blog/artificial-intelligence-ai-and-chatgpt-history-and-timelines 16 | - text: Timeline of AI and language models 17 | url: https://lifearchitect.ai/timeline/ 18 | - text: The Road to AGI 2015 - 2025 19 | url: https://ai-timeline.org/ 20 | - text: AI 2027 21 | url: https://ai-2027.com -------------------------------------------------------------------------------- /feed.xml: -------------------------------------------------------------------------------- 1 | --- 2 | layout: none 3 | --- 4 | 5 | 6 | 7 | AI Timeline RSS Feed 8 | {{ site.url | default: "/" | absolute_url }}{{ site.baseurl }} 9 | A comprehensive timeline of Artificial Intelligence milestones from 2022 to present. 10 | 11 | {% for year in site.data.timeline reversed %} 12 | {% for event in year.events reversed %} 13 | {% assign date_str = "1 " | append: event.date | append: " " | append: year.year %} 14 | 15 | {{ event.date }} {{ year.year }} 16 | {{ site.url | default: "/" | absolute_url }}{{ site.baseurl }}/#{{ year.year }}-{{ event.date | replace: ' ', '-' }} 17 | {{ site.url | default: "/" | absolute_url }}{{ site.baseurl }}/#{{ year.year }}-{{ event.date | replace: ' ', '-' }} 18 | {{ date_str | date: "%a, %d %b %Y %H:%M:%S %z" }} 19 | 21 | {% for info in event.info %} 22 |
  • {{ info.text }}
    • 23 | {% endfor %} 24 | 25 | ]]>     26 |     27 | {% endfor %} 28 | {% endfor %} 29 |     30 |     31 | -------------------------------------------------------------------------------- /Gemfile.lock: -------------------------------------------------------------------------------- 1 | GEM 2 | remote: https://rubygems.org/ 3 | specs: 4 | addressable (2.8.7) 5 | public_suffix (>= 2.0.2, < 7.0) 6 | base64 (0.3.0) 7 | bigdecimal (3.1.8) 8 | colorator (1.1.0) 9 | concurrent-ruby (1.3.4) 10 | csv (3.3.5) 11 | em-websocket (0.5.3) 12 | eventmachine (>= 0.12.9) 13 | http_parser.rb (~> 0) 14 | eventmachine (1.2.7) 15 | ffi (1.17.2-x64-mingw-ucrt) 16 | ffi (1.17.2-x86_64-linux-gnu) 17 | forwardable-extended (2.6.0) 18 | google-protobuf (4.32.0-x64-mingw-ucrt) 19 | bigdecimal 20 | rake (>= 13) 21 | google-protobuf (4.32.0-x86_64-linux-gnu) 22 | bigdecimal 23 | rake (>= 13) 24 | http_parser.rb (0.8.0) 25 | i18n (1.14.7) 26 | concurrent-ruby (~> 1.0) 27 | jekyll (4.4.1) 28 | addressable (~> 2.4) 29 | base64 (~> 0.2) 30 | colorator (~> 1.0) 31 | csv (~> 3.0) 32 | em-websocket (~> 0.5) 33 | i18n (~> 1.0) 34 | jekyll-sass-converter (>= 2.0, < 4.0) 35 | jekyll-watch (~> 2.0) 36 | json (~> 2.6) 37 | kramdown (~> 2.3, >= 2.3.1) 38 | kramdown-parser-gfm (~> 1.0) 39 | liquid (~> 4.0) 40 | mercenary (~> 0.3, >= 0.3.6) 41 | pathutil (~> 0.9) 42 | rouge (>= 3.0, < 5.0) 43 | safe_yaml (~> 1.0) 44 | terminal-table (>= 1.8, < 4.0) 45 | webrick (~> 1.7) 46 | jekyll-sass-converter (3.1.0) 47 | sass-embedded (~> 1.75) 48 | jekyll-watch (2.2.1) 49 | listen (~> 3.0) 50 | json (2.13.2) 51 | kramdown (2.5.1) 52 | rexml (>= 3.3.9) 53 | kramdown-parser-gfm (1.1.0) 54 | kramdown (~> 2.0) 55 | liquid (4.0.4) 56 | listen (3.9.0) 57 | rb-fsevent (~> 0.10, >= 0.10.3) 58 | rb-inotify (~> 0.9, >= 0.9.10) 59 | logger (1.7.0) 60 | mercenary (0.4.0) 61 | pathutil (0.16.2) 62 | forwardable-extended (~> 2.6) 63 | public_suffix (6.0.2) 64 | rake (13.3.0) 65 | rb-fsevent (0.11.2) 66 | rb-inotify (0.11.1) 67 | ffi (~> 1.0) 68 | rexml (3.4.1) 69 | rouge (4.6.0) 70 | safe_yaml (1.0.5) 71 | sass-embedded (1.90.0-x64-mingw-ucrt) 72 | google-protobuf (~> 4.31) 73 | sass-embedded (1.90.0-x86_64-linux-gnu) 74 | google-protobuf (~> 4.31) 75 | terminal-table (3.0.2) 76 | unicode-display_width (>= 1.1.1, < 3) 77 | unicode-display_width (2.6.0) 78 | webrick (1.9.1) 79 | 80 | PLATFORMS 81 | x64-mingw-ucrt 82 | x86_64-linux 83 | 84 | DEPENDENCIES 85 | base64 86 | csv 87 | jekyll 88 | logger 89 | 90 | BUNDLED WITH 91 | 2.6.9 92 | -------------------------------------------------------------------------------- /.github/workflows/deploy.yml: -------------------------------------------------------------------------------- 1 | # Sample workflow for building and deploying a Jekyll site to GitHub Pages 2 | name: Deploy Jekyll site to Pages 3 | 4 | on: 5 | # Runs on pushes targeting the default branch 6 | push: 7 | branches: ["main"] 8 | 9 | # Allows you to run this workflow manually from the Actions tab 10 | workflow_dispatch: 11 | 12 | # Sets permissions of the GITHUB_TOKEN to allow deployment to GitHub Pages 13 | permissions: 14 | contents: write 15 | pages: write 16 | id-token: write 17 | 18 | # Allow only one concurrent deployment, skipping runs queued between the run in-progress and latest queued. 19 | concurrency: 20 | group: "pages" 21 | cancel-in-progress: false 22 | 23 | jobs: 24 | # Build job 25 | build: 26 | runs-on: ubuntu-latest 27 | steps: 28 | - name: Checkout 29 | uses: actions/checkout@v4 30 | 31 | - name: Setup Python 32 | uses: actions/setup-python@v5 33 | with: 34 | python-version: '3.11' 35 | 36 | - name: Install Python dependencies 37 | run: | 38 | python -m pip install --upgrade pip 39 | pip install -r requirements.txt 40 | 41 | - name: Convert timeline.md to timeline.yml 42 | run: python scripts/convert_timeline_events.py _data/timeline.md 43 | 44 | - name: Commit converted timeline.yml 45 | uses: stefanzweifel/git-auto-commit-action@v5 46 | with: 47 | commit_message: "Automated: Convert timeline.md to timeline.yml" 48 | file_pattern: _data/timeline.yml 49 | 50 | - name: Setup Ruby 51 | uses: ruby/setup-ruby@v1 52 | with: 53 | ruby-version: '3.4' 54 | 55 | - name: Add linux platform to bundle 56 | run: bundle lock --add-platform x86_64-linux 57 | 58 | - name: Install dependencies 59 | run: bundle install --jobs 4 --retry 3 60 | 61 | - name: Setup Pages 62 | id: pages 63 | uses: actions/configure-pages@v5 64 | 65 | - name: Build with Jekyll 66 | run: bundle exec jekyll build --baseurl "${{ steps.pages.outputs.base_path }}" 67 | env: 68 | JEKYLL_ENV: production 69 | 70 | - name: Upload artifact 71 | uses: actions/upload-pages-artifact@v3 72 | 73 | # Deployment job 74 | deploy: 75 | environment: 76 | name: github-pages 77 | url: ${{ steps.deployment.outputs.page_url }} 78 | runs-on: ubuntu-latest 79 | needs: build 80 | steps: 81 | - name: Deploy to GitHub Pages 82 | id: deployment 83 | uses: actions/deploy-pages@v4 -------------------------------------------------------------------------------- /_sass/_variables.scss: -------------------------------------------------------------------------------- 1 | :root { 2 | /* --- Colors --- */ 3 | --bg-color: #ffffff; 4 | --text-color: #1d1d1f; 5 | --text-secondary: #86868b; 6 | --border-color: #e5e5ea; 7 | 8 | /* Interactive */ 9 | --accent-color: #0066cc; 10 | --accent-hover: #004499; 11 | --accent-bg-subtle: rgba(0, 102, 204, 0.1); 12 | 13 | /* Elements */ 14 | --milestone-color: #f5a623; 15 | --card-bg: #f5f5f7; 16 | --sticky-bg: rgba(255, 255, 255, 0.95); 17 | --footer-bg: #f5f5f7; 18 | --hover-bg: #f5f5f7; 19 | --input-bg: #f5f5f7; 20 | 21 | /* --- Typography --- */ 22 | --font-family-base: -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, Helvetica, Arial, sans-serif; 23 | 24 | --font-size-xs: 12px; 25 | --font-size-sm: 14px; 26 | --font-size-base: 16px; 27 | --font-size-lg: 18px; 28 | --font-size-xl: 24px; 29 | --font-size-2xl: 32px; 30 | --font-size-3xl: 48px; 31 | 32 | --font-weight-regular: 400; 33 | --font-weight-medium: 500; 34 | --font-weight-semibold: 600; 35 | --font-weight-bold: 700; 36 | --font-weight-extra-bold: 800; 37 | 38 | /* --- Spacing --- */ 39 | --space-1: 4px; 40 | --space-2: 8px; 41 | --space-3: 12px; 42 | --space-4: 16px; 43 | --space-5: 20px; 44 | --space-6: 24px; 45 | --space-8: 32px; 46 | --space-10: 40px; 47 | --space-12: 50px; 48 | --space-15: 60px; 49 | --space-20: 80px; 50 | 51 | /* --- Layout --- */ 52 | --container-width: 900px; 53 | --timeline-width-narrow: 760px; 54 | --header-scroll-padding: 155px; /* Offset for sticky header */ 55 | --timeline-line-x: 29px; 56 | --timeline-stroke-width: 2px; 57 | 58 | /* --- Borders & Radius --- */ 59 | --radius-sm: 8px; 60 | --radius-md: 12px; 61 | --radius-pill: 99px; 62 | 63 | /* --- Effects --- */ 64 | --shadow-sm: 0 2px 8px rgba(0,0,0,0.04); 65 | --shadow-md: 0 4px 12px rgba(0,0,0,0.08); 66 | --shadow-lg: 0 8px 24px rgba(0,0,0,0.12); 67 | --shadow-hover: 0 4px 15px rgba(0,0,0,0.1); 68 | 69 | --transition-fast: 0.2s ease; 70 | --transition-base: 0.3s ease; 71 | --backdrop-blur: 20px; 72 | 73 | /* --- Z-Index --- */ 74 | --z-back: 0; 75 | --z-content: 2; 76 | --z-sticky: 900; 77 | --z-tooltip: 1000; 78 | } 79 | 80 | body.dark-mode { 81 | --bg-color: #121212; 82 | --text-color: #f5f5f7; 83 | --text-secondary: #a1a1a6; 84 | --border-color: #2c2c2e; 85 | 86 | --accent-color: #409cff; 87 | --accent-hover: #0077ed; 88 | --accent-bg-subtle: rgba(64, 156, 255, 0.15); 89 | 90 | --milestone-color: #ffd60a; 91 | 92 | --card-bg: #1c1c1e; 93 | --sticky-bg: rgba(18, 18, 18, 0.95); 94 | --footer-bg: #1c1c1e; 95 | --hover-bg: #2c2c2e; 96 | --input-bg: #1c1c1e; 97 | 98 | --shadow-sm: 0 2px 8px rgba(0,0,0,0.2); 99 | --shadow-md: 0 4px 12px rgba(0,0,0,0.3); 100 | --shadow-hover: 0 4px 15px rgba(0,0,0,0.4); 101 | } 102 | -------------------------------------------------------------------------------- /index.md: -------------------------------------------------------------------------------- 1 | --- 2 | layout: default 3 | title: AI Timeline 4 | description: A comprehensive timeline of Artificial Intelligence milestones from 2022 to present. 5 | --- 6 | 7 |
    8 |

    AI Timeline

    9 |

    Tracking the evolution of Artificial Intelligence

    10 | 11 | 12 | Star on GitHub 13 | 14 |
    15 | 16 |
    17 | 18 |
    19 |
    20 | 21 | 22 |
    23 |
    24 | 25 | 26 |
    27 | 32 | 33 |
    34 | 37 | 38 | 41 | 42 | 45 |
    46 |
    47 |
    48 | 49 |
    50 | 51 | {% for year in site.data.timeline reversed %} 52 |
    53 |

    54 | {{ year.year }} 55 | 58 |

    59 | {% for event in year.events reversed %} 60 | {% assign event_id = year.year | append: '-' | append: event.date | replace: ' ', '-' %} 61 |
    62 |
    63 | {{ event.date | date: "%B" }} 64 | 67 |
    68 | 69 | {% for info in event.info %} 70 | {% if info.special %} 71 |
    72 | {{ info.text }} 73 |
    74 | {% else %} 75 |
    76 | {{ info.text }} 77 |
    78 | {% endif %} 79 | {% endfor %} 80 |
    81 | {% endfor %} 82 |
    83 | {% endfor %} 84 | 85 |
    86 | 87 |
    88 |

    Enrichment Resources

    89 | 99 |
    100 | 101 | 104 | 105 | 108 | 109 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # AI Timeline 2 | 3 | An open-source timeline tracking the evolution and advancements in Artificial Intelligence from 2022 onwards. 4 | 5 | ## Overview 6 | 7 | This project provides a chronological overview of significant milestones, model releases, and key developments in the AI field. It aims to be a clear and concise resource for tracking the industry's rapid progress. 8 | 9 | ## Features 10 | 11 | - **Yearly Sections:** Events organized by year. 12 | - **Monthly Events:** Clear breakdown of events within each month. 13 | - **Key Event Highlighting:** Important milestones are visually emphasized. 14 | - **Sortable View:** Toggle between newest-first and oldest-first chronological order. 15 | - **Dark Mode:** Switch between light and dark themes. 16 | - **Responsive Design:** Works well on desktop and mobile devices. 17 | - **Further Reading:** Links to related articles and resources. 18 | 19 | ## How to Use 20 | 21 | 1. **Visit the Timeline:** [https://nhlocal.github.io/AiTimeline/](https://nhlocal.github.io/AiTimeline/) 22 | 2. **Navigate:** Scroll or use the year links at the top. 23 | 3. **Sort:** Use the "Sort" button to change the timeline order. 24 | 4. **Theme:** Use the toggle button (moon/sun icon) to switch between dark and light modes. 25 | 26 | ## How the Site Works 27 | 28 | This site is built using **Jekyll**, a static site generator, and is automatically deployed using **GitHub Actions**. 29 | 30 | - **Content Source:** The single source of truth for the timeline is `_data/timeline.md`. This is the only file you need to edit to add or change events. 31 | - **Data File:** The `_data/timeline.yml` file is used by Jekyll to build the site. **This file is generated automatically** by a script. Do not edit it directly. 32 | - **Automation:** When changes are pushed to the `main` branch, a GitHub Action automatically runs: 33 | 1. Converts the `_data/timeline.md` file to `_data/timeline.yml`. 34 | 2. Commits the updated `.yml` file back to the repository. 35 | 3. Builds the Jekyll site. 36 | 4. Deploys the site to GitHub Pages. 37 | 38 | ## Contributing 39 | 40 | Contributions are welcome and have been simplified! Help keep the timeline accurate and up-to-date. 41 | 42 | 1. **Fork & Clone:** Fork the repository to your GitHub account and clone it locally. 43 | ```bash 44 | git clone https://github.com/YOUR_USERNAME/AiTimeline.git 45 | cd AiTimeline 46 | ``` 47 | 48 | 2. **Edit `_data/timeline.md`:** 49 | * Open the `_data/timeline.md` file. 50 | * Add or modify events using the existing format (`# Year:`, `## Month`, `- Event description`). 51 | * Use `**bold text**` for emphasis (e.g., model names). 52 | * To highlight an event as "special", add `(*special*)` at the end of its line, with a space before it. 53 | ```markdown 54 | ## March 55 | - An important event happened. (*special*) 56 | - A regular event occurred. 57 | ``` 58 | 59 | 3. **Commit and Push Your Changes:** 60 | * Stage and commit **only the `timeline.md` file**. The `.yml` file will be updated automatically by the workflow. 61 | ```bash 62 | git add _data/timeline.md 63 | git commit -m "docs: Add [brief description of your change]" 64 | ``` 65 | * Push your changes to your forked repository: 66 | ```bash 67 | git push origin main # or your branch name 68 | ``` 69 | 70 | 4. **Create a Pull Request:** 71 | * Go to your fork on GitHub and open a Pull Request to the `main` branch of `NHLOCAL/AiTimeline`. 72 | * Once your Pull Request is merged, the automated workflow will handle the conversion and deployment. 73 | 74 | ## Feedback and Suggestions 75 | 76 | Have feedback, found a bug, or want to suggest an addition? Please **[open an issue](https://github.com/NHLOCAL/AiTimeline/issues)** on GitHub. We appreciate your input! 77 | 78 | ## Running Locally (Optional) 79 | 80 | To preview your changes on your local machine before creating a pull request: 81 | 82 | 1. **Install Prerequisites:** 83 | * Ruby and Bundler (see [Jekyll Installation Guide](https://jekyllrb.com/docs/installation/)) 84 | * Python 3 85 | 86 | 2. **Install Dependencies:** 87 | ```bash 88 | bundle install 89 | ``` 90 | 91 | 3. **Generate the `.yml` File for Local Preview:** 92 | * After editing `_data/timeline.md`, you must run the conversion script to see your changes locally. 93 | * This step is **only for local previewing**. The final conversion is handled automatically online. 94 | ```bash 95 | # Run from the project root 96 | python scripts/convert_timeline_events.py _data/timeline.md 97 | ``` 98 | 99 | 4. **Serve the Site:** 100 | ```bash 101 | bundle exec jekyll serve 102 | ``` 103 | 104 | 5. **View:** Open your browser to `http://localhost:4000/AiTimeline/` (or the address provided). 105 | 106 | ## Star History 107 | 108 | [![Star History Chart](https://api.star-history.com/svg?repos=NHLOCAL/AiTimeline&type=Date)](https://www.star-history.com/#NHLOCAL/AiTimeline&Date) 109 | 110 | ## License 111 | 112 | ### Code License 113 | 114 | The code in this repository is licensed under the MIT License. See the [LICENSE](LICENSE) file for details. 115 | 116 | ### Content License 117 | 118 | The content of the website, including the timeline data, is licensed under the [Creative Commons Attribution 4.0 International License](http://creativecommons.org/licenses/by/4.0/). 119 | 120 | This means you are free to share and adapt the content for any purpose, even commercially, as long as you give appropriate credit to the original source. 121 | 122 | ### How to Attribute 123 | 124 | If you use or reference the data from this timeline, please use one of the following standard citation formats. 125 | 126 | **APA Style:** 127 | ``` 128 | NHLOCAL. (2025). *AI Timeline*. Retrieved from https://nhlocal.github.io/AiTimeline/ 129 | ``` 130 | 131 | **BibTeX Entry:** 132 | ```bibtex 133 | @misc{nhlocal_ai_timeline, 134 | author = {NHLOCAL}, 135 | title = {AI Timeline}, 136 | year = {2025}, 137 | howpublished = {\url{https://nhlocal.github.io/AiTimeline/}}, 138 | } 139 | ``` -------------------------------------------------------------------------------- /notes.md: -------------------------------------------------------------------------------- 1 | ## August 2025 2 | - Google introduced **Gemini 2.5 Deep Think**, a special "extended thinking" mode for solving complex problems and exploring alternatives. (*special*) 3 | - Anthropic released **Claude Opus 4.1**, an upgrade focused on improving agentic capabilities and real-world coding. 4 | - Google DeepMind announced **Genie 3.0**, a "world model" for creating interactive 3D environments from text, maintaining consistency for several minutes. (*special*) 5 | - OpenAI released **gpt-oss-120b** and **gpt-oss-20b**, a family of open-source models with high reasoning capabilities, optimized to run on accessible hardware. 6 | - OpenAI launched **GPT-5**, the company's next-generation model, with significant improvements in coding and a dynamic "thinking" mode to reduce hallucinations. (*special*) 7 | - DeepSeek released **DeepSeek V3.1**, a hybrid model combining fast and slow "thinking" modes to improve performance in agentic tasks and tool use. 8 | - Google launched a preview of **Gemini 2.5 Flash Image** (showcased as *nano-banana*), an advanced model for precise image editing, merging, and maintaining character consistency. (*special*) 9 | 10 | ## ספטמבר 2025 11 | - Seedream 4.0 של ByteDance הוא מודל דור חדש ליצירת ועריכת תמונות במהירות ובאיכות גבוהה (עד 4K), המשלב יצירה ועריכה במערכת אחת ומתחרה ישירות במובילי השוק כמו Google DeepMind. 12 | - Deep Think achieves Gold Medal at the ICPC 2025 Programming Contest 13 | - OpenAI Reasoning Model Solved ALL 12 Problems at ICPC 2025 Programming Contest 14 | - suno V5 שוחרר 15 | - qwen 3 max 16 | - wan 2.2 - מודל וידאו גנרטיבי מתקדם המסוגל לערוך והלחליף דמות לאורך סרט שלם 17 | - אנתרופיק משחררת את מודל קלוד סונט 4.5 18 | - openAI משיקה את SORA 2 19 | 20 | 21 | ## September 2025 22 | 23 | - ByteDance released **Seedream 4.0**, a next-generation image model unifying high-quality text-to-image generation and natural-language image editing. 24 | - An advanced Gemini variant, reported as **Gemini 2.5 - Deep Think**, achieved gold-medal-level performance at the ICPC World Finals programming contest. 25 | - OpenAI reported a reasoning and code model achieved a **perfect score (12/12)** in ICPC testing. 26 | - Suno released **Suno v5**, an upgrade in music generation with studio-grade fidelity and more natural-sounding vocals. 27 | - Alibaba unveiled **Qwen-3-Max**, its flagship model with over a trillion parameters, focusing on long context and agent capabilities. 28 | - **Wan 2.2** was released, a generative video model focused on multi-shot consistency and character animation. 29 | - Anthropic announced **Claude Sonnet 4.5**, a model optimized for coding, agent construction, and improved reasoning. 30 | - OpenAI released **Sora 2**, a flagship video and audio generation model with improved physical modeling and synchronized sound. 31 | - DeepSeek released **DeepSeek-V3.2-Exp** 32 | - OpenAI and NVIDIA announced a strategic partnership for NVIDIA to supply at least **10 gigawatts** of AI systems for OpenAI's infrastructure. 33 | 34 | 35 | ## אוקטובר 2025 36 | - FIGURE 03 37 | - מודל SOTA לשליטה במחשב מבוסס ג'מיני שוחרר על ידי גוגל 38 | - הייקו 4.5 - ביצועים ואיכות גבוהה מאוד בעלות מינימלית 39 | - צ'אט GPT אטלס - דפדפן בינה מלאכותית 40 | - הרובוט ההומנואידי הראשון לשימוש ביתי מסחרי הושק - 1X Neo 41 | 42 | ## October 2025 43 | 44 | - Figure unveiled **Figure 03**, a humanoid robot designed for domestic and general-purpose tasks. 45 | - Google released a **Gemini model for computer control**, achieving state-of-the-art (SOTA) performance in GUI automation. 46 | - Anthropic released **Claude 4.5 Haiku**, a fast, cost-effective model for high-volume, low-latency applications. 47 | - OpenAI announced **ChatGPT Atlas**, an AI-native web browser with a built-in "Agent Mode" for task automation. 48 | - 1X announced **Neo**, a humanoid robot marketed as the first consumer-ready model for home use. 49 | 50 | 51 | ## נובמבר 2025 52 | - KINI K2 THINKING שוחרר והגיע לראש המדדים 53 | - GPT 5.1 54 | - GROK 4.1 55 | - GEMINI 3.0 56 | - GPT 5.1 CODEX MAX 57 | - ננו בננה פרו (2) 58 | - קלוד אופוס 4.5 59 | - FLUX 2 - מודל יצירת תמונת קוד פתוח 60 | - DeepseekMath-V2 מקבל את מדליית הזהב באולימפיאדת המתמטיקה 61 | - מיקרוסופט משחררת את Fara-7B כקוד פתוח - מודל שליטה במחשב 62 | 63 | 64 | ## November 2025 65 | 66 | - Moonshot AI released **Kimi K2 Thinking**, an open model setting new records in reasoning benchmarks. 67 | - OpenAI launched **GPT 5.1**, featuring specialized "Thinking" and "Instant" modes with expanded context. 68 | - xAI released **Grok 4.1**, combining high EQ with strong logic to top the LM Arena leaderboard. 69 | - Google debuted **Gemini 3.0**, a flagship "thinking" model that claimed the top spot on major benchmarks. 70 | - OpenAI introduced **GPT 5.1 Codex Max**, an agentic model built specifically for long-term coding tasks. 71 | - Google released **Nano Banana Pro**, a superior image generation and editing model based on Gemini 3. 72 | - Anthropic announced **Claude Opus 4.5**, delivering elite coding and agentic performance at a significantly reduced price. 73 | - Black Forest Labs launched **FLUX 2**, a high-performance open-weight image generation model. 74 | - DeepSeek released **DeepSeekMath-V2** as open source, achieving gold-medal performance in math olympiads. 75 | - Microsoft open-sourced **Fara-7B**, a small model optimized for browser agents and computer control. 76 | 77 | 78 | ## נובמבר 2025 79 | 80 | - חברת Moonshot AI שחררה את **Kimi K2 Thinking**, מודל פתוח שקבע שיאים חדשים במדדי הסקה. 81 | - חברת OpenAI השיקה את **GPT 5.1**, הכולל מצבי "Thinking" ו-"Instant" עם חלון הקשר מורחב. 82 | - חברת xAI שחררה את **Grok 4.1**, המשלב EQ גבוה עם לוגיקה חזקה והוביל את טבלת ה-LM Arena. 83 | - גוגל הציגה את **Gemini 3.0**, מודל חשיבה שתפס את המקום הראשון במדדים המובילים. 84 | - חברת OpenAI הציגה את **GPT 5.1 Codex Max**, מודל סוכן שנבנה במיוחד למשימות קוד ארוכות. 85 | - גוגל שחררה את **Nano Banana Pro**, מודל עריכה ויצירת תמונות מתקדם המבוסס על Gemini 3. 86 | - חברת Anthropic הכריזה על **Claude Opus 4.5**, המציע יכולות קוד וסוכנים מובילות במחיר מופחת משמעותית. 87 | - מעבדות Black Forest Labs שחררו את **FLUX 2**, מודל תמונה בקוד פתוח עם ביצועים גבוהים. 88 | - חברת DeepSeek שחררה את **DeepSeekMath-V2** כקוד פתוח, שהשיג ביצועי מדליית זהב באולימפיאדות מתמטיקה. 89 | - מיקרוסופט שחררה כקוד פתוח את **Fara-7B**, מודל קטן המותאם לסוכני דפדפן ושליטה במחשב. -------------------------------------------------------------------------------- /scripts/convert_timeline_events.py: -------------------------------------------------------------------------------- 1 | import yaml 2 | import re 3 | import os 4 | import sys 5 | 6 | def md_to_yaml(md_content): 7 | all_years_data = [] # <--- List to store structures for ALL years 8 | current_year_structure = None # Holds data for the year currently being processed 9 | current_event = None # Holds data for the event currently being processed 10 | 11 | lines = md_content.splitlines() 12 | 13 | for line in lines: 14 | line = line.strip() 15 | if not line: 16 | continue # Skip empty lines 17 | 18 | if line.startswith("# Year: "): 19 | # 1. Finalize and store the PREVIOUS year's structure (if one exists) 20 | if current_year_structure: 21 | if current_event: # Add the last event of the previous year 22 | current_year_structure["events"].append(current_event) 23 | current_event = None # Reset event since it's now stored 24 | all_years_data.append(current_year_structure) # Add completed year to the list 25 | 26 | # 2. Start the NEW year's structure 27 | try: 28 | year_num = int(line.replace("# Year: ", "")) 29 | current_year_structure = {"year": year_num, "events": []} 30 | current_event = None # Ensure event is reset for the new year 31 | except ValueError: 32 | print(f"Warning: Could not parse year from line: '{line}'. Skipping year.") 33 | current_year_structure = None # Invalidate current year structure 34 | 35 | elif line.startswith("## "): 36 | # Ensure we are inside a valid year structure 37 | if not current_year_structure: 38 | print(f"Warning: Found month line '{line}' outside of a year (# Year:). Skipping.") 39 | continue 40 | 41 | # Finalize the previous event within the CURRENT year 42 | if current_event: 43 | current_year_structure["events"].append(current_event) 44 | 45 | # Start the new event for the CURRENT year 46 | event_date = line.replace("## ", "") 47 | current_event = {"date": event_date, "info": []} 48 | 49 | elif line.startswith("- "): 50 | # Ensure we are inside a valid event structure 51 | if not current_event: 52 | # We could also check current_year_structure here for extra safety 53 | print(f"Warning: Found info line '{line}' outside of an event (## Month). Skipping.") 54 | continue 55 | 56 | # Process the info line for the current event 57 | text = line.replace("- ", "", 1).strip() 58 | is_special = False 59 | if text.endswith("(*special*)"): 60 | text = text[:-len("(*special*)")].strip() # Remove the marker 61 | is_special = True 62 | 63 | # Convert markdown bold to HTML bold 64 | # Use raw string r"\1" for replacement to avoid issues with backslashes 65 | text = re.sub(r"\*\*(.*?)\*\*", r"\1", text) 66 | 67 | info_entry = {"text": text} 68 | if is_special: 69 | info_entry["special"] = True # Add the special flag only if needed 70 | 71 | current_event["info"].append(info_entry) 72 | 73 | if current_year_structure: 74 | if current_event: # Add the last event of the last year 75 | current_year_structure["events"].append(current_event) 76 | all_years_data.append(current_year_structure) # Add the final year to the list 77 | 78 | if not all_years_data: 79 | print("Warning: No year data found or processed.") 80 | return "" # Return empty string or handle as error 81 | 82 | return yaml.safe_dump(all_years_data, sort_keys=False, allow_unicode=True, default_flow_style=False, indent=2, width=float("inf")) 83 | 84 | def yaml_to_md(yaml_content): 85 | data = yaml.safe_load(yaml_content) 86 | md_lines = [] 87 | 88 | if not isinstance(data, list): 89 | print("Warning: Expected YAML input to be a list of years.") 90 | if isinstance(data, dict) and "year" in data: 91 | data = [data] 92 | else: 93 | return "Error: Invalid YAML format." 94 | 95 | for item in data: 96 | if "year" in item and item["year"]: 97 | md_lines.append(f"# Year: {item['year']}") 98 | md_lines.append("") 99 | 100 | for event in item.get("events", []): 101 | md_lines.append(f"## {event['date']}") 102 | for info in event.get("info", []): 103 | text = info.get("text", "") 104 | is_special = info.get("special", False) 105 | 106 | text = re.sub(r"(.*?)", r"**\1**", text) 107 | 108 | if is_special: 109 | md_lines.append(f"- {text} (*special*)") 110 | else: 111 | md_lines.append(f"- {text}") 112 | md_lines.append("") 113 | md_lines.append("") 114 | 115 | while md_lines and not md_lines[-1]: 116 | md_lines.pop() 117 | 118 | return "\n".join(md_lines) 119 | 120 | # Main function to handle file inputs 121 | def main(): 122 | if len(sys.argv) < 2: 123 | print("Error: Please provide an input file path as a command-line argument.") 124 | print(f"Usage: python {sys.argv[0]} ") 125 | sys.exit(1) 126 | 127 | input_file = sys.argv[1] 128 | 129 | if not os.path.exists(input_file): 130 | print(f"Error: Input file '{input_file}' does not exist.") 131 | return 132 | 133 | file_extension = os.path.splitext(input_file)[1].lower() 134 | output_file = "" 135 | 136 | if file_extension == ".md": 137 | conversion_type = "md-to-yaml" 138 | output_file = input_file.replace(".md", ".yml") 139 | print(f"Converting Markdown ('{input_file}') to YAML ('{output_file}')...") 140 | elif file_extension in [".yaml", ".yml"]: 141 | conversion_type = "yaml-to-md" 142 | output_file = input_file.replace(file_extension, ".md") 143 | print(f"Converting YAML ('{input_file}') to Markdown ('{output_file}')...") 144 | else: 145 | print("Error: Unsupported file format. Please use .md or .yaml/.yml files.") 146 | return 147 | 148 | try: 149 | with open(input_file, "r", encoding="utf-8") as file: 150 | content = file.read() 151 | except Exception as e: 152 | print(f"Error reading input file '{input_file}': {e}") 153 | return 154 | 155 | converted_content = "" 156 | try: 157 | if conversion_type == "md-to-yaml": 158 | converted_content = md_to_yaml(content) 159 | elif conversion_type == "yaml-to-md": 160 | converted_content = yaml_to_md(content) 161 | except Exception as e: 162 | print(f"Error during conversion: {e}") 163 | return 164 | 165 | if not converted_content or converted_content.startswith("Error:"): 166 | print(f"Conversion failed. No output written.") 167 | if converted_content: print(converted_content) 168 | return 169 | 170 | try: 171 | with open(output_file, "w", encoding="utf-8") as file: 172 | file.write(converted_content) 173 | print(f"Conversion completed successfully. Output written to '{output_file}'.") 174 | except Exception as e: 175 | print(f"Error writing output file '{output_file}': {e}") 176 | 177 | 178 | if __name__ == "__main__": 179 | main() -------------------------------------------------------------------------------- /assets/js/script.js: -------------------------------------------------------------------------------- 1 | const darkModeToggle = document.getElementById('dark-mode-toggle'); 2 | const body = document.body; 3 | const icon = darkModeToggle ? darkModeToggle.querySelector('i') : null; 4 | 5 | 6 | function setCookie(name, value, days) { 7 | let expires = ""; 8 | if (days) { 9 | const date = new Date(); 10 | date.setTime(date.getTime() + (days * 24 * 60 * 60 * 1000)); 11 | expires = "; expires=" + date.toUTCString(); 12 | } 13 | document.cookie = name + "=" + (value || "") + expires + "; path=/; SameSite=Lax"; 14 | } 15 | 16 | function getCookie(name) { 17 | const nameEQ = name + "="; 18 | const ca = document.cookie.split(';'); 19 | for (let i = 0; i < ca.length; i++) { 20 | let c = ca[i]; 21 | while (c.charAt(0) === ' ') c = c.substring(1, c.length); 22 | if (c.indexOf(nameEQ) === 0) return c.substring(nameEQ.length, c.length); 23 | } 24 | return null; 25 | } 26 | 27 | function updateIcon(isDark) { 28 | if (!icon) return; 29 | icon.className = isDark ? 'fas fa-sun' : 'fas fa-moon'; 30 | } 31 | 32 | function initTheme() { 33 | const savedTheme = getCookie("darkMode"); 34 | const systemPrefersDark = window.matchMedia("(prefers-color-scheme: dark)").matches; 35 | 36 | if (savedTheme === "true") { 37 | body.classList.add('dark-mode'); 38 | updateIcon(true); 39 | } else if (savedTheme === "false") { 40 | body.classList.remove('dark-mode'); 41 | updateIcon(false); 42 | } else { 43 | if (systemPrefersDark) { 44 | body.classList.add('dark-mode'); 45 | updateIcon(true); 46 | } else { 47 | body.classList.remove('dark-mode'); 48 | updateIcon(false); 49 | } 50 | } 51 | } 52 | 53 | if (darkModeToggle) { 54 | initTheme(); 55 | darkModeToggle.addEventListener('click', () => { 56 | body.classList.toggle('dark-mode'); 57 | const isDark = body.classList.contains('dark-mode'); 58 | updateIcon(isDark); 59 | setCookie("darkMode", isDark, 365); 60 | }); 61 | } 62 | 63 | 64 | const yearNavLinks = document.querySelectorAll('.year-nav a'); 65 | 66 | function highlightNav() { 67 | if (yearNavLinks.length === 0) return; 68 | 69 | const currentSections = Array.from(document.querySelectorAll('.year:not(.hidden)')); 70 | let currentYear = ''; 71 | 72 | // Adjust offset to match CSS scroll-margin-top (~155px) 73 | // We add a bit more (e.g. 180) so highlighting happens when the year header is well into view 74 | const scrollPaddingTop = parseInt(getComputedStyle(document.documentElement).getPropertyValue('--header-scroll-padding'), 10) || 155; 75 | const scrollPos = window.pageYOffset + scrollPaddingTop + 25; 76 | 77 | currentSections.forEach(section => { 78 | if (scrollPos >= section.offsetTop) { 79 | currentYear = section.getAttribute('id'); 80 | } 81 | }); 82 | 83 | if (!currentYear && currentSections.length > 0) { 84 | currentYear = currentSections[0].getAttribute('id'); 85 | } 86 | 87 | yearNavLinks.forEach(link => { 88 | link.classList.remove('active'); 89 | if (link.getAttribute('href') === '#' + currentYear) { 90 | link.classList.add('active'); 91 | 92 | 93 | } 94 | }); 95 | } 96 | 97 | window.addEventListener('scroll', highlightNav); 98 | window.addEventListener('load', highlightNav); 99 | 100 | 101 | const sortToggle = document.getElementById('sort-toggle'); 102 | if (sortToggle) { 103 | sortToggle.addEventListener('click', () => { 104 | const timelineContainer = document.querySelector('.timeline'); 105 | const currentOrder = sortToggle.getAttribute('data-order') || 'newest'; 106 | const newOrder = currentOrder === 'oldest' ? 'newest' : 'oldest'; 107 | 108 | sortToggle.setAttribute('data-order', newOrder); 109 | const icon = sortToggle.querySelector('i'); 110 | 111 | if (newOrder === 'newest') { 112 | icon.className = 'fas fa-sort-amount-up'; 113 | } else { 114 | icon.className = 'fas fa-sort-amount-down'; 115 | } 116 | 117 | const yearSectionsArray = Array.from(timelineContainer.querySelectorAll('.year')); 118 | yearSectionsArray.reverse().forEach(section => { 119 | timelineContainer.appendChild(section); 120 | const events = Array.from(section.querySelectorAll('.event')); 121 | events.reverse().forEach(event => { 122 | section.appendChild(event); 123 | }); 124 | }); 125 | 126 | highlightNav(); 127 | }); 128 | } 129 | 130 | 131 | 132 | const searchInput = document.getElementById('event-search'); 133 | const significantFilter = document.getElementById('filter-significant'); 134 | 135 | 136 | let activeFilters = { 137 | search: '', 138 | special: false 139 | }; 140 | 141 | // Initialize: Cache original HTML content for each info item to allow safe highlight reset 142 | const allInfoItems = document.querySelectorAll('.info'); 143 | allInfoItems.forEach(item => { 144 | item.setAttribute('data-original-html', item.innerHTML); 145 | }); 146 | 147 | function escapeRegExp(string) { 148 | return string.replace(/[.*+?^${}()|[\]\\]/g, '\\$&'); 149 | } 150 | 151 | function filterEvents() { 152 | const years = document.querySelectorAll('.year'); 153 | 154 | years.forEach(year => { 155 | let hasVisibleEventsInYear = false; 156 | const events = year.querySelectorAll('.event'); 157 | 158 | events.forEach(event => { 159 | let hasVisibleItemsInEvent = false; 160 | const items = event.querySelectorAll('.info'); 161 | 162 | items.forEach(item => { 163 | 164 | const originalHTML = item.getAttribute('data-original-html'); 165 | 166 | 167 | const isSpecial = item.getAttribute('data-special') === 'true'; 168 | const textContent = item.innerText.toLowerCase(); 169 | const searchMatch = activeFilters.search === '' || textContent.includes(activeFilters.search); 170 | const specialMatch = !activeFilters.special || isSpecial; 171 | 172 | if (searchMatch && specialMatch) { 173 | item.classList.remove('hidden'); 174 | hasVisibleItemsInEvent = true; 175 | 176 | 177 | if (activeFilters.search !== '') { 178 | // Safe highlight: Match text not inside HTML tags 179 | try { 180 | const term = escapeRegExp(activeFilters.search); 181 | // Regex looks for the term, ensuring it's not followed by `>` without a `<` first (rudimentary tag avoidance) 182 | 183 | const regex = new RegExp(`(${term})(?![^<]*>)`, 'gi'); 184 | item.innerHTML = originalHTML.replace(regex, '$1'); 185 | } catch (e) { 186 | item.innerHTML = originalHTML; 187 | } 188 | } else { 189 | item.innerHTML = originalHTML; 190 | } 191 | 192 | } else { 193 | item.classList.add('hidden'); 194 | 195 | item.innerHTML = originalHTML; 196 | } 197 | }); 198 | 199 | 200 | if (hasVisibleItemsInEvent) { 201 | event.classList.remove('hidden'); 202 | hasVisibleEventsInYear = true; 203 | } else { 204 | event.classList.add('hidden'); 205 | } 206 | }); 207 | 208 | 209 | if (hasVisibleEventsInYear) { 210 | year.classList.remove('hidden'); 211 | } else { 212 | year.classList.add('hidden'); 213 | } 214 | }); 215 | 216 | highlightNav(); 217 | } 218 | 219 | if (searchInput) { 220 | // Stop browser defaults (history back / page scroll) when user hits delete/backspace on an empty search box. 221 | searchInput.addEventListener('keydown', (e) => { 222 | const isEraseKey = e.key === 'Backspace' || e.key === 'Delete'; 223 | if (!isEraseKey) return; 224 | 225 | const value = searchInput.value; 226 | const isEmpty = value.length === 0; 227 | // Only block when there is truly nothing to delete; allow normal deletion when text exists (even if all selected) 228 | if (!isEmpty) return; 229 | 230 | e.preventDefault(); 231 | e.stopPropagation(); 232 | }); 233 | 234 | searchInput.addEventListener('input', (e) => { 235 | const next = e.target.value.toLowerCase().trim(); 236 | if (next === activeFilters.search) return; // no content change -> avoid unnecessary reflow 237 | activeFilters.search = next; 238 | filterEvents(); 239 | }); 240 | } 241 | 242 | if (significantFilter) { 243 | significantFilter.addEventListener('click', () => { 244 | activeFilters.special = !activeFilters.special; 245 | 246 | const icon = significantFilter.querySelector('i'); 247 | if (activeFilters.special) { 248 | significantFilter.classList.add('active'); 249 | icon.className = 'fas fa-star'; 250 | } else { 251 | significantFilter.classList.remove('active'); 252 | icon.className = 'far fa-star'; 253 | } 254 | 255 | filterEvents(); 256 | }); 257 | } 258 | 259 | 260 | const scrollToTopBtn = document.getElementById('scroll-to-top'); 261 | 262 | if (scrollToTopBtn) { 263 | window.addEventListener('scroll', () => { 264 | if (window.pageYOffset > 500) { 265 | scrollToTopBtn.classList.add('visible'); 266 | } else { 267 | scrollToTopBtn.classList.remove('visible'); 268 | } 269 | }); 270 | 271 | scrollToTopBtn.addEventListener('click', () => { 272 | window.scrollTo({ 273 | top: 0, 274 | behavior: 'smooth' 275 | }); 276 | }); 277 | } 278 | 279 | 280 | document.addEventListener('click', function(e) { 281 | const btn = e.target.closest('.anchor-btn'); 282 | if (!btn) return; 283 | 284 | const linkId = btn.getAttribute('data-link'); 285 | if (!linkId) return; 286 | 287 | const fullUrl = window.location.origin + window.location.pathname + '#' + linkId; 288 | 289 | navigator.clipboard.writeText(fullUrl).then(() => { 290 | const icon = btn.querySelector('i'); 291 | const originalClass = icon.className; 292 | 293 | icon.className = 'fas fa-check'; 294 | icon.style.color = 'var(--milestone-color)'; 295 | 296 | setTimeout(() => { 297 | icon.className = originalClass; 298 | icon.style.color = ''; 299 | }, 1500); 300 | }).catch(err => { 301 | console.error('Failed to copy: ', err); 302 | }); 303 | }); 304 | -------------------------------------------------------------------------------- /_sass/_layout.scss: -------------------------------------------------------------------------------- 1 | html { 2 | /* Smooth scrolling caused visible jumps when the layout changed; disable globally */ 3 | scroll-behavior: auto; 4 | scroll-padding-top: var(--header-scroll-padding); 5 | /* Stop root scroll anchoring from jumping when timeline length changes */ 6 | overflow-anchor: none; 7 | } 8 | 9 | body { 10 | font-family: var(--font-family-base); 11 | margin: 0; 12 | padding: 0; 13 | line-height: 1.6; 14 | color: var(--text-color); 15 | background-color: var(--bg-color); 16 | transition: background-color var(--transition-base), color var(--transition-base); 17 | /* Prevent browser scroll anchoring jumps when search hides/shows items */ 18 | overflow-anchor: none; 19 | } 20 | 21 | 22 | .header { 23 | text-align: center; 24 | padding: var(--space-20) var(--space-5) var(--space-12); 25 | background-color: var(--bg-color); 26 | 27 | h1 { 28 | font-size: var(--font-size-3xl); 29 | font-weight: var(--font-weight-bold); 30 | margin: 0 0 var(--space-3); 31 | letter-spacing: -0.02em; 32 | line-height: 1.1; 33 | } 34 | 35 | h2 { 36 | font-size: var(--font-size-xl); 37 | font-weight: var(--font-weight-regular); 38 | color: var(--text-secondary); 39 | margin: 0 0 35px; 40 | } 41 | } 42 | 43 | 44 | .github-cta { 45 | display: inline-flex; 46 | align-items: center; 47 | gap: var(--space-2); 48 | background-color: var(--text-color); 49 | color: var(--bg-color); 50 | padding: 10px var(--space-6); 51 | border-radius: var(--radius-pill); 52 | text-decoration: none; 53 | font-weight: var(--font-weight-semibold); 54 | font-size: 15px; 55 | transition: all var(--transition-fast); 56 | border: 1px solid transparent; 57 | 58 | &:hover { 59 | transform: translateY(-2px); 60 | opacity: 0.9; 61 | text-decoration: none; 62 | color: var(--bg-color); 63 | box-shadow: var(--shadow-hover); 64 | } 65 | 66 | 67 | &.small { 68 | font-size: var(--font-size-sm); 69 | padding: 6px 14px; 70 | } 71 | } 72 | 73 | 74 | .sticky-bar { 75 | position: sticky; 76 | top: 0; 77 | z-index: var(--z-sticky); 78 | background-color: var(--sticky-bg); 79 | backdrop-filter: blur(var(--backdrop-blur)); 80 | -webkit-backdrop-filter: blur(var(--backdrop-blur)); 81 | border-bottom: 1px solid var(--border-color); 82 | padding: var(--space-4) var(--space-5); 83 | display: flex; 84 | flex-direction: column; 85 | gap: var(--space-4); 86 | transition: padding var(--transition-base); 87 | box-shadow: var(--shadow-sm); 88 | } 89 | 90 | .bar-row { 91 | display: flex; 92 | justify-content: space-between; 93 | align-items: center; 94 | gap: var(--space-5); 95 | width: 100%; 96 | max-width: var(--container-width); 97 | margin: 0 auto; 98 | } 99 | 100 | 101 | .search-container { 102 | width: 100%; 103 | max-width: 600px; 104 | margin: 0 auto; 105 | } 106 | 107 | .search-wrapper { 108 | position: relative; 109 | display: flex; 110 | align-items: center; 111 | color: var(--text-secondary); 112 | background-color: var(--input-bg); 113 | border: 1px solid transparent; 114 | border-radius: var(--radius-pill); 115 | padding: 0 var(--space-5); 116 | transition: all var(--transition-fast); 117 | height: 48px; 118 | width: 100%; 119 | box-sizing: border-box; 120 | 121 | &:hover { 122 | background-color: var(--hover-bg); 123 | } 124 | 125 | &:focus-within { 126 | background-color: var(--bg-color); 127 | border-color: var(--accent-color); 128 | box-shadow: 0 0 0 4px var(--accent-bg-subtle); 129 | color: var(--text-color); 130 | } 131 | 132 | i { 133 | font-size: var(--font-size-base); 134 | margin-right: var(--space-3); 135 | opacity: 0.7; 136 | } 137 | 138 | input { 139 | background: transparent; 140 | border: none; 141 | color: var(--text-color); 142 | font-family: inherit; 143 | font-size: var(--font-size-base); 144 | padding: var(--space-2) 0; 145 | width: 100%; 146 | outline: none; 147 | 148 | &::placeholder { 149 | color: var(--text-secondary); 150 | opacity: 0.7; 151 | } 152 | } 153 | } 154 | 155 | 156 | .year-nav { 157 | display: flex; 158 | gap: var(--space-1); 159 | flex-wrap: wrap; 160 | align-items: center; 161 | 162 | a { 163 | padding: 6px 14px; 164 | border-radius: var(--radius-pill); 165 | text-decoration: none; 166 | color: var(--text-secondary); 167 | font-weight: var(--font-weight-medium); 168 | font-size: var(--font-size-sm); 169 | transition: all var(--transition-fast); 170 | background-color: transparent; 171 | 172 | &:hover { 173 | color: var(--text-color); 174 | background-color: var(--hover-bg); 175 | } 176 | 177 | &.active { 178 | color: var(--bg-color); 179 | background-color: var(--text-color); 180 | font-weight: var(--font-weight-semibold); 181 | } 182 | } 183 | } 184 | 185 | 186 | .controls-container { 187 | display: flex; 188 | gap: var(--space-2); 189 | align-items: center; 190 | 191 | .control-btn { 192 | background: transparent; 193 | border: none; 194 | cursor: pointer; 195 | color: var(--text-secondary); 196 | font-size: var(--font-size-sm); 197 | height: 36px; 198 | padding: 0 var(--space-3); 199 | display: inline-flex; 200 | align-items: center; 201 | justify-content: center; 202 | gap: var(--space-2); 203 | font-family: inherit; 204 | font-weight: var(--font-weight-medium); 205 | transition: all var(--transition-fast); 206 | border-radius: var(--radius-sm); 207 | 208 | span { 209 | display: none; 210 | } 211 | 212 | &:hover { 213 | background-color: var(--hover-bg); 214 | color: var(--text-color); 215 | } 216 | 217 | &.active { 218 | color: var(--accent-color); 219 | background-color: var(--accent-bg-subtle); 220 | font-weight: var(--font-weight-semibold); 221 | } 222 | } 223 | } 224 | 225 | @media (min-width: 768px) { 226 | .controls-container .control-btn span { 227 | display: inline; 228 | } 229 | } 230 | 231 | 232 | .timeline { 233 | max-width: var(--timeline-width-narrow); 234 | margin: 0 auto; 235 | padding: var(--space-15) var(--space-5) 0; 236 | position: relative; 237 | /* Avoid scroll-anchoring jumps when search hides/shows content */ 238 | overflow-anchor: none; 239 | 240 | &::before { 241 | content: ''; 242 | position: absolute; 243 | top: 0; 244 | bottom: 0; 245 | left: var(--timeline-line-x); 246 | width: var(--timeline-stroke-width); 247 | background-color: var(--border-color); 248 | z-index: var(--z-back); 249 | } 250 | } 251 | 252 | .year { 253 | margin-bottom: var(--space-20); 254 | position: relative; 255 | scroll-margin-top: var(--header-scroll-padding); 256 | overflow-anchor: none; 257 | 258 | &.hidden { 259 | display: none; 260 | } 261 | 262 | h2 { 263 | font-size: var(--font-size-2xl); 264 | font-weight: var(--font-weight-extra-bold); 265 | color: var(--text-color); 266 | 267 | background-color: var(--sticky-bg); 268 | backdrop-filter: blur(var(--backdrop-blur)); 269 | -webkit-backdrop-filter: blur(var(--backdrop-blur)); 270 | border-bottom: 1px solid var(--border-color); 271 | width: 100%; 272 | margin: 0 0 var(--space-10) 0; 273 | padding: var(--space-3) 0; 274 | 275 | position: sticky; 276 | top: calc(var(--header-scroll-padding) - 20px); 277 | z-index: 10; 278 | line-height: 1; 279 | 280 | display: flex; 281 | align-items: center; 282 | gap: var(--space-3); 283 | 284 | &:hover .anchor-btn { 285 | opacity: 1; 286 | } 287 | } 288 | } 289 | 290 | .anchor-btn { 291 | background: none; 292 | border: none; 293 | color: var(--text-secondary); 294 | cursor: pointer; 295 | font-size: var(--font-size-sm); 296 | padding: var(--space-2); 297 | opacity: 0.3; 298 | transition: all var(--transition-fast); 299 | display: inline-flex; 300 | align-items: center; 301 | justify-content: center; 302 | 303 | &:focus, &:hover { 304 | opacity: 1 !important; 305 | color: var(--accent-color); 306 | transform: scale(1.1); 307 | } 308 | } 309 | 310 | .event { 311 | margin-bottom: var(--space-12); 312 | padding-left: 60px; 313 | padding-top: var(--space-4); 314 | padding-bottom: var(--space-4); 315 | position: relative; 316 | scroll-margin-top: var(--header-scroll-padding); 317 | overflow-anchor: none; 318 | border-radius: var(--radius-md); 319 | 320 | border: 1px solid transparent; 321 | border-left: 4px solid transparent; 322 | transition: all 0.2s cubic-bezier(0.25, 1, 0.5, 1); 323 | 324 | &.hidden { 325 | display: none; 326 | } 327 | 328 | &:hover { 329 | background-color: var(--card-bg); 330 | border-color: var(--border-color); 331 | border-left-color: var(--accent-color); 332 | transform: translateX(5px); 333 | } 334 | 335 | .date { 336 | font-size: 13px; 337 | text-transform: uppercase; 338 | letter-spacing: 1.5px; 339 | font-weight: var(--font-weight-bold); 340 | color: var(--accent-color); 341 | margin-bottom: var(--space-4); 342 | display: inline-flex; 343 | align-items: center; 344 | gap: var(--space-2); 345 | border-bottom: 1px solid transparent; 346 | transition: border-color var(--transition-fast); 347 | 348 | &:hover { 349 | border-bottom-color: var(--accent-color); 350 | 351 | .anchor-btn { 352 | opacity: 1; 353 | } 354 | } 355 | 356 | .anchor-btn { 357 | font-size: 12px; 358 | } 359 | } 360 | 361 | .info { 362 | position: relative; 363 | margin-bottom: var(--space-4); 364 | font-size: 17px; 365 | color: var(--text-color); 366 | line-height: 1.6; 367 | 368 | &.hidden { 369 | display: none; 370 | } 371 | 372 | /* Timeline Dots */ 373 | &::before { 374 | content: ''; 375 | position: absolute; 376 | top: 9px; 377 | left: calc(var(--timeline-line-x) - 60px + (var(--timeline-stroke-width) / 2)); 378 | transform: translateX(-50%); 379 | width: 9px; 380 | height: 9px; 381 | background-color: var(--bg-color); 382 | border: 2px solid var(--border-color); 383 | border-radius: 50%; 384 | z-index: var(--z-content); 385 | transition: all 0.2s cubic-bezier(0.25, 1, 0.5, 1); 386 | box-shadow: 0 0 0 2px var(--bg-color); 387 | } 388 | 389 | &.special { 390 | /* Star Icon */ 391 | &::before { 392 | content: '\f005'; 393 | font-family: "Font Awesome 6 Free"; 394 | font-weight: 900; 395 | font-size: var(--font-size-xs); 396 | line-height: 1; 397 | width: auto; 398 | height: auto; 399 | border: none; 400 | border-radius: 0; 401 | top: 7px; 402 | color: var(--milestone-color); 403 | background-color: transparent; 404 | padding: 2px 0; 405 | box-shadow: none; 406 | text-shadow: 0 0 10px rgba(var(--milestone-color), 0.3); 407 | } 408 | } 409 | } 410 | 411 | &:hover .info::before { 412 | border-color: var(--accent-color); 413 | background-color: var(--bg-color); 414 | transform: translateX(-50%) scale(1.2); 415 | } 416 | 417 | &:hover .info.special::before { 418 | background-color: transparent; 419 | transform: translateX(-50%) scale(1.3) rotate(15deg); 420 | box-shadow: none; 421 | text-shadow: 0 0 15px rgba(var(--milestone-color), 0.6); 422 | } 423 | } 424 | 425 | 426 | .highlight-text { 427 | background-color: rgba(255, 212, 0, 0.3); 428 | color: var(--text-color); 429 | border-radius: 2px; 430 | padding: 0 2px; 431 | } 432 | body.dark-mode .highlight-text { 433 | background-color: rgba(255, 212, 0, 0.25); 434 | color: #fff; 435 | } 436 | 437 | 438 | .resources-section { 439 | max-width: 800px; 440 | margin: 100px auto 60px; 441 | padding: 0 var(--space-5); 442 | position: relative; 443 | 444 | &::before { 445 | content: ''; 446 | display: block; 447 | width: 100%; 448 | height: 1px; 449 | background: linear-gradient(90deg, transparent, var(--border-color), transparent); 450 | margin-bottom: 50px; 451 | } 452 | 453 | h3 { 454 | font-size: 13px; 455 | font-weight: var(--font-weight-bold); 456 | text-transform: uppercase; 457 | letter-spacing: 2px; 458 | color: var(--text-secondary); 459 | text-align: center; 460 | margin-bottom: var(--space-10); 461 | position: relative; 462 | display: flex; 463 | align-items: center; 464 | justify-content: center; 465 | gap: var(--space-4); 466 | 467 | &::after, &::before { 468 | content: ''; 469 | width: 40px; 470 | height: 2px; 471 | background-color: var(--accent-color); 472 | border-radius: 2px; 473 | } 474 | } 475 | } 476 | 477 | .resources-list { 478 | list-style: none; 479 | padding: 0; 480 | margin: 0; 481 | display: grid; 482 | grid-template-columns: 1fr; 483 | gap: var(--space-4); 484 | 485 | li { 486 | border-bottom: none; 487 | } 488 | 489 | a { 490 | display: flex; 491 | justify-content: space-between; 492 | align-items: center; 493 | padding: var(--space-5) 25px; 494 | text-decoration: none; 495 | color: var(--text-color); 496 | background-color: var(--card-bg); 497 | border: 1px solid transparent; 498 | border-radius: var(--radius-md); 499 | transition: all var(--transition-fast); 500 | font-weight: var(--font-weight-medium); 501 | 502 | &:hover { 503 | transform: translateY(-2px); 504 | background-color: var(--bg-color); 505 | border-color: var(--border-color); 506 | box-shadow: var(--shadow-hover); 507 | color: var(--accent-color); 508 | 509 | .resource-icon { 510 | opacity: 1; 511 | transform: translateX(0); 512 | } 513 | } 514 | } 515 | } 516 | 517 | .resource-text { 518 | font-size: var(--font-size-base); 519 | } 520 | 521 | .resource-icon { 522 | font-size: var(--font-size-sm); 523 | opacity: 0; 524 | transform: translateX(-10px); 525 | transition: all var(--transition-fast); 526 | color: var(--accent-color); 527 | } 528 | 529 | 530 | 531 | footer { 532 | background-color: var(--footer-bg); 533 | border-top: 1px solid var(--border-color); 534 | padding: var(--space-20) 0 60px; 535 | margin-top: 0; 536 | font-size: var(--font-size-sm); 537 | } 538 | 539 | .footer-container { 540 | max-width: var(--container-width); 541 | margin: 0 auto; 542 | padding: 0 var(--space-5); 543 | display: flex; 544 | justify-content: space-between; 545 | flex-wrap: wrap; 546 | gap: 50px; 547 | } 548 | 549 | .footer-col { 550 | flex: 1; 551 | min-width: 250px; 552 | 553 | h3 { 554 | font-size: 12px; 555 | text-transform: uppercase; 556 | letter-spacing: 1px; 557 | color: var(--text-secondary); 558 | margin-bottom: var(--space-5); 559 | font-weight: var(--font-weight-bold); 560 | } 561 | 562 | p { 563 | color: var(--text-secondary); 564 | margin-bottom: var(--space-5); 565 | max-width: 320px; 566 | line-height: 1.6; 567 | } 568 | } 569 | 570 | .footer-links { 571 | list-style: none; 572 | padding: 0; 573 | margin: 0; 574 | 575 | li { margin-bottom: var(--space-3); } 576 | 577 | a { 578 | color: var(--text-color); 579 | text-decoration: none; 580 | display: inline-flex; 581 | align-items: center; 582 | gap: 10px; 583 | opacity: 0.8; 584 | transition: all var(--transition-fast); 585 | font-weight: var(--font-weight-medium); 586 | 587 | &:hover { 588 | opacity: 1; 589 | color: var(--accent-color); 590 | transform: translateX(3px); 591 | } 592 | 593 | i { width: 16px; text-align: center; } 594 | } 595 | } 596 | 597 | 598 | .scroll-to-top { 599 | position: fixed; 600 | bottom: 30px; 601 | right: 30px; 602 | width: 48px; 603 | height: 48px; 604 | background-color: var(--bg-color); 605 | color: var(--text-color); 606 | border: 1px solid var(--border-color); 607 | border-radius: 50%; 608 | cursor: pointer; 609 | display: flex; 610 | justify-content: center; 611 | align-items: center; 612 | opacity: 0; 613 | visibility: hidden; 614 | transform: translateY(20px); 615 | transition: all var(--transition-base); 616 | z-index: var(--z-tooltip); 617 | box-shadow: var(--shadow-md); 618 | 619 | &.visible { 620 | opacity: 1; 621 | visibility: visible; 622 | transform: translateY(0); 623 | } 624 | 625 | &:hover { 626 | transform: translateY(-3px); 627 | background-color: var(--accent-color); 628 | color: #fff; 629 | border-color: var(--accent-color); 630 | } 631 | } 632 | 633 | 634 | @media (max-width: 768px) { 635 | :root { 636 | --timeline-line-x: 18px; 637 | } 638 | 639 | .header { 640 | padding: 50px var(--space-5) 30px; 641 | h1 { font-size: 36px; } 642 | h2 { font-size: 18px; margin-bottom: 25px; } 643 | } 644 | 645 | .sticky-bar { 646 | padding: var(--space-3) 15px; 647 | gap: var(--space-3); 648 | } 649 | 650 | .bar-row { 651 | flex-direction: column-reverse; 652 | gap: var(--space-3); 653 | } 654 | 655 | .year-nav { 656 | width: 100%; 657 | overflow-x: auto; 658 | white-space: nowrap; 659 | justify-content: center; 660 | padding-bottom: 5px; 661 | scrollbar-width: none; 662 | &::-webkit-scrollbar { display: none; } 663 | } 664 | 665 | .timeline { padding-top: 30px; } 666 | 667 | .year h2 { 668 | top: 170px; 669 | font-size: 26px; 670 | padding-top: 15px; 671 | } 672 | 673 | .event { 674 | padding-left: 40px; 675 | 676 | .info::before, .info.special::before { 677 | left: calc(var(--timeline-line-x) - 40px + (var(--timeline-stroke-width) / 2)); 678 | } 679 | } 680 | 681 | .footer-container { 682 | flex-direction: column; 683 | gap: 40px; 684 | } 685 | 686 | .resources-section h3 { 687 | gap: 10px; 688 | &::before, &::after { width: 25px; } 689 | } 690 | } 691 | -------------------------------------------------------------------------------- /_data/timeline.md: -------------------------------------------------------------------------------- 1 | # Year: 2022 2 | 3 | ## February 4 | - **Midjourney v1** 5 | 6 | ## March 7 | - OpenAI releases **text-davinci-002** and **code-davinci-002** with an API approach. 8 | 9 | ## April 10 | - **Midjourney v2** 11 | - **DALL-E 2** is announced for gradual release. (*special*) 12 | 13 | ## July 14 | - **Midjourney v3** is launched. 15 | 16 | ## August 17 | - **Stable Diffusion 1.4** is released. 18 | 19 | ## October 20 | - **Stable Diffusion 1.5** becomes available. (*special*) 21 | 22 | ## November 23 | - **ChatGPT**, a chatbot by OpenAI using GPT-3.5, is released to the public and quickly becomes a viral sensation. (*special*) 24 | - **Midjourney v4** is released. 25 | - **Stable Diffusion 2.0** is launched. 26 | 27 | ## December 28 | - **Stable Diffusion 2.1** is released. 29 | 30 | 31 | # Year: 2023 32 | 33 | ## February 34 | - Meta releases the **LLaMA** language model as open-source for research purposes. The model is later leaked. (*special*) 35 | - Microsoft gradually releases **Bing AI**, an AI chat based on an upgraded GPT model integrating internet search. 36 | 37 | ## March 38 | - **Midjourney v5** is launched. 39 | - OpenAI's **GPT-4** model is partially released, featuring multimodal image analysis and improved multi-language support. (*special*) 40 | - Google releases the AI chat **Bard** in a limited capacity, based on the LaMDA language model. 41 | 42 | ## April 43 | - Adobe releases the **Firefly** image creation model as a beta version to a waiting list. The model allowed a variety of capabilities including text formatting. 44 | 45 | ## May 46 | - **Midjourney v5.1** is released. 47 | - Google announces an upgrade to Bard, moving it to the upgraded **PaLM 2** language model. It will support 180 countries and many languages. 48 | 49 | ## June 50 | - **Midjourney v5.2** is launched. 51 | 52 | ## July 53 | - **Stable Diffusion XL 1.0** is released. 54 | - Anthropic announces a new version of their large language model - **Claude 2**. 55 | - Meta releases the **LLaMA 2** open source language model to the general public in a variety of sizes. 56 | 57 | ## October 58 | - **DALL-E 3** is released. 59 | - Adobe releases **Firefly 2**. 60 | 61 | ## November 62 | - **Stable Diffusion XL Turbo** is released - A fast model that allows the creation of an image in one step in real-time. 63 | 64 | ## December 65 | - **Midjourney v6** is launched. 66 | - Google upgrades Bard in limited areas, moving it to be based on the upgraded **Gemini Pro** language model. 67 | - X Corporation launches **Grok AI** chatbot for paid subscribers in English language. 68 | 69 | 70 | # Year: 2024 71 | 72 | ## February 73 | - Stability AI announces **Stable Diffusion 3** (gradually released to waiting list). 74 | - Google upgrades the artificial intelligence chat in Bard, basing it on the new **Gemini Pro** model, in all available languages. Google replaces "Bard" with "Gemini". 75 | - Google announces the **Gemini Pro 1.5** multimodal language model capable of parsing up to a million tokens, as well as parsing video and images. The model is gradually released to developers on a waiting list. (*special*) 76 | - OpenAI announces the **Sora** model that produces videos up to a minute long. The model is not released to the public at this time. (*special*) 77 | 78 | ## March 79 | - X Corporation announces the upcoming release of the **Grok 1.5** open source model. 80 | - Anthropic announces **Claude 3**, a new version of their large language model. The version is deployed in 3 different sizes, with the largest model performing better than GPT-4. 81 | - Suno AI, which develops a model for creating music, releases **Suno v3** to the general public. 82 | 83 | ## April 84 | - Stability AI releases a new update to the music creation model - **Stable Audio 2.0**. 85 | - X Corporation releases an upgrade to its language model, **Grok-1.5V**, which integrates high-level image recognition. In the test presented by the company, the model is the best in identifying and analyzing images compared to other models. 86 | - The Mistral company releases its new model **Mixtral 8x22B** as open source. This is the most powerful model among the open source models and it contains 141 billion parameters but uses a method that allows more economical use. 87 | - Meta releases the **LLaMA 3** model as open source in sizes 8B and 70B parameters. The large model shows better performance than Claude 3 Sonnet and Gemini Pro 1.5 in several measures. Meta is expected to later release larger models with 400 billion parameters and more. 88 | - Microsoft releases the **Phi-3-mini** model in open source. The model comes in a reduced version of 3.8B parameters, which allows it to run on mobile devices as well, and it presents capabilities similar to GPT-3.5. (*special*) 89 | - Adobe announces its new image creation model **Firefly 3**. 90 | - The startup **Reka AI** presents a series of multimodal language models in 3 sizes. The models are capable of processing video, audio and images. The large model featured similar capabilities to GPT-4. 91 | - Apple releases as full open source a series of small language models under the name **OpenELM**. The models are available in four weights between 270 million and 3 billion parameters. 92 | 93 | ## May 94 | - OpenAI announces the **GPT-4o model** that presents full multimodal capabilities, including receiving and creating text, images, and audio. The model presents an impressive ability to speak with a high response speed and in natural language. The model is 2 times more efficient than the GPT-4 Turbo model, and has better capabilities for languages other than English. (*special*) 95 | - Google announces a large number of AI features in its products. The main ones: increasing the token limit to 2 million for Gemini 1.5 to waiting list, releasing a smaller and faster **Gemini Flash 1.5 model**. Revealing the latest image creation model **Imagen 3**, music creation model **Music AI** and video creation model **Veo**. And the announcement of the **Astra model** with multimodal capabilities for realtime audio and video reception. 96 | - Microsoft announces **Copilot+** for dedicated computers, which will allow a full search of the user's history through screenshots of the user's activity. The company also released as open source the SLMs that display impressive capabilities in a minimal size: **Phi-3 Small**, **Phi-3 Medium**, and **Phi-3 Vision** which includes image recognition capability. 97 | - Meta introduces **Chameleon**, a new multimodal model that seamlessly renders text and images. 98 | - Mistral AI releases a new open source version of its language model **Mistral-7B-Instruct-v0.3**. 99 | - Google announces **AI Overviews** intended to give a summary of the relevant information in Google search. (*special*) 100 | - Suno AI releases an updated music creation model **Suno v3.5**. 101 | - Mistral AI releases a new language model designed for coding **Codestral** in size 22B. 102 | 103 | ## June 104 | - Stability AI releases its updated image creation model **Stable Diffusion 3** in a medium version in size 2B parameters. 105 | - Apple announces **Apple Intelligence**, an AI system that will be integrated into the company's devices and will combine AI models of different sizes for different tasks. 106 | - DeepSeekAI publishes the **DeepSeekCoderV2** open source language model which presents similar coding capabilities to models such as GPT-4, Claude 3 Opus and more. 107 | - **Runway** introduces **Gen3 Alpha**, a new AI model for video generation. 108 | - Anthropic releases the **Claude Sonnet 3.5** model, which presents better capabilities than other models with low resource usage. (*special*) 109 | - Microsoft releases in open source a series of image recognition models called **Florence 2**. 110 | - Google announces **Gemma 2** open source language models with 9B and 27B parameter sizes. Also, the company opens the context window capabilities to developers for up to 2 million tokens. 111 | 112 | ## July 113 | - OpenAI has released a miniaturized model called **GPT-4o mini** that presents high capabilities at a low cost 114 | - Meta releases as open source the **llama 3.1 model** in sizes 8B, 70B and 405B. The large model features the same capabilities as the best closed source models (*special*) 115 | - mistral ai releases three new models: **Codestral Mamba**, **Mistral NeMo** and **Mathstral** designed for mathematics 116 | - Google DeepMind has unveiled two new AI systems that won silver medals at this year's International Mathematical Olympiad (IMO), **AlphaProof** and **AlphaGeometry 2**. (*special*) 117 | - OpenAI launched **SearchGPT**, an integrated web search 118 | - Startup Udio has released **Udio v1.5**, an updated version of its music creation model 119 | - Mistral AI has released a large language model **Mistral Large 2** in size 123B, which presents capabilities close to the closed SOTA models. (*special*) 120 | - **Midjourney v6.1** is released 121 | - Google releases the **Gemma 2 2B** model as open source. The model demonstrates better capabilities than much larger models. 122 | 123 | ## August 124 | - "Black Forest Labs" releases weights for an image creation model named **Flux**, which shows better performance than similar closedsource models. 125 | - OpenAI released a new version of its model, **GPT-4o 0806**, achieving 100% success in generating valid JSON output. 126 | - Google's image generation model, **Imagen 3**, has been released. 127 | - xAI Corporation has launched the models **Grok 2** and **Grok 2 mini**, which demonstrate performance on par with leading SOTA models in the market. 128 | - Microsoft has introduced its small language models, **Phi 3.5**, in three versions, each showcasing impressive performance relative to their size. 129 | - Google has introduced three new experimental AI models: **Gemini 1.5 Flash8B**, **Gemini 1.5 Pro** Enhanced, and **Gemini 1.5 Flash** Updated. 130 | - **Ideogram 2.0** has been released, offering image generation capabilities that surpass those of other leading models. 131 | - Luma has unveiled the **Dream Machine 1.5** model for video creation. 132 | 133 | ## September 134 | - The French AI company Mistral has introduced **Pixtral12B**, its first multimodal model capable of processing both images and text. 135 | - OPENAI has released two nextgeneration AI models to its subscribers: **o1 preview** and **o1 mini**. These models show a significant improvement in performance, particularly in tasks requiring reasoning, including coding, mathematics, GPQA, and more. (*special*) 136 | - Chinese company Alibaba releases the **Qwen 2.5** model in various sizes, ranging from 0.5B to 72B. The models demonstrate capabilities comparable to much larger models. 137 | - The video generation model **KLING 1.5** has been released. 138 | - **OpenAI** launches the **advanced voice mode** of GPT4o for all subscribers. 139 | - **Meta** releases **Llama 3.2** in sizes 1B, 3B, 11B and 90B, featuring image recognition capabilities for the first time. 140 | - **Google** has rolled out new model updates ready for deployment, **Gemini Pro 1.5 002** and **Gemini Flash 1.5 002**, showcasing significantly improved longcontext processing. 141 | - **Kyutai** releases two opensource versions of its voicetovoice model, **Moshi**. 142 | - Google releases an update to its AI tool **NotebookLM** that enables users to create podcasts based on their own content. 143 | - Mistral AI launches a 22B model named **Mistral Small**. 144 | 145 | ## October 146 | - **Flux 1.1 Pro** is released, showcasing advanced capabilities for image creation. 147 | - Meta unveils **Movie Gen**, a new AI model that generates videos, images, and audio from text input. 148 | - Pika introduces **Video Model 1.5** along with "Pika Effects." 149 | - Adobe announces its video creation model, **Firefly Video**. 150 | - Startup Rhymes AI releases **Aria**, an opensource, multimodal model exhibiting capabilities similar to comparably sized proprietary models. 151 | - Meta releases an opensource speechtospeech language model named **Meta Spirit LM**. 152 | - Mistral AI introduces **Ministral**, a new model available in 3B and 8B parameter sizes. 153 | - **Janus AI**, a multimodal language model capable of recognizing and generating both text and images, is released as open source by DeepSeekAI. 154 | - Google DeepMind and MIT unveil **Fluid**, a texttoimage generation model with industryleading performance at a scale of 10.5B parameters. 155 | - **Stable Diffusion 3.5** is released in three sizes as open source. 156 | - Anthropic launches **Claude 3.5 Sonnet New**, demonstrating significant advancements in specific areas over its previous version, and announces **Claude 3.5 Haiku**. 157 | - Anthropic announces an experimental feature for computer use with a public beta API. 158 | - The texttoimage model **Recraft v3** has been released to the public, ranking first in benchmarks compared to similar models. 159 | - OpenAI has launched **Search GPT**, allowing users to perform web searches directly within the platform. 160 | 161 | ## November 162 | - Alibaba released its new model, **QwQ 32B Preview**, which integrates reasoning capabilities before responding. The model competes with, and sometimes surpasses, OpenAI's o1-preview model. 163 | - Alibaba opensourced the model **Qwen2.5 Coder 32B**, which offers comparable capabilities to leading proprietary language models in the coding domain. 164 | - DeepSeek unveiled its new AI model, **DeepSeek-R1-Lite-Preview**, which incorporates reasoning capabilities and delivers impressive performance on the AIME and MATH benchmarks, matching the level of OpenAI's o1-preview. 165 | - **Suno** upgraded its AIpowered music generator to **v4**, introducing new features and performance improvements. 166 | - Mistral AI launched the **Pixtral Large** model, a multimodal language model excelling in image recognition and advanced performance metrics, and an update to Mistral Large, 2411. 167 | - Google introduced two experimental models, **gemini-exp-1114** and **gemini-exp-1121**, currently leading the arena chatbot with enhanced performance. 168 | - Anthropic launches **Claude 3.5 Haiku** and Visual PDF Analysis in Claude. 169 | 170 | ## December 171 | - Amazon introduced a new series of models called **NOVA**, designed for text, image, and video processing. 172 | - OpenAI released **SORA**, a video generation model, along with the full version of **O1** and **O1 Pro** for advanced subscribers. Additionally, the company launched a live video mode for **GPT4o**. (*special*) 173 | - Google unveiled the experimental model **Gemini-Exp-1206**, which ranked first in the chatbot leaderboard. 174 | - Google launched **Gemini 2.0 Flash** in beta. This model leads benchmarks and outperforms the previous version, **Gemini Pro 1.5**. Additionally, Google introduced live speech and video mode and announced built-in image generation capabilities within the model. (*special*) 175 | - Google revealed **Gemini-2.0-Flash-Thinking**, a thinking model based on **Gemini 2.0 Flash**, which secured second place in the chatbot leaderboard. (*special*) 176 | - Google introduced **Veo 2**, a beta version video generation model capable of producing 4K videos up to two minutes long. The model outperformed **SORA** in human evaluations. Additionally, Google updated **Imagen 3**, offering enhanced image quality and realism. (*special*) 177 | - xAI integrated **Aurora**, a new model for generating high-quality and realistic images. 178 | - Microsoft open-sourced the **Phi4** model, sized at 14B, showcasing impressive capabilities for its size. 179 | - Meta released **Llama 3.3 70B**, a model offering performance comparable to **Llama 3.1 405B**. 180 | - Google launched a multi-modal open-source model called **PaliGemma 2**, integrated with existing **Gemma** models. 181 | - Pika Labs released **2.0**, the latest version of its AI-powered video generator. 182 | - Meta introduced **Apollo**, a video generation model available in three different sizes. 183 | - Deepseek open-sourced **Deepseek V3**, a model with 671B parameters that surpasses closed-source SOTA models across several benchmarks. (*special*) 184 | - Alibaba unveiled **QVQ-72B-Preview**, a cutting-edge thinking model capable of analyzing images, featuring SOTA-level performance. (*special*) 185 | - OpenAI announced **O3**, a groundbreaking AI model achieving 87.5% in the **ARC-AGI** benchmark, 25.2% in the **Frontier Math Benchmark** (compared to under 2% in previous models), and 87.7% in Ph.D.-level science questions. A cost-effective version, **O3 Mini**, is expected in January 2025, with performance similar to **O1**, alongside improved speed and efficiency. (*special*) 186 | - The video generation model **Kling 1.6** was released, offering significant performance enhancements. 187 | 188 | 189 | # Year: 2025 190 | 191 | ## January 192 | - OpenAI released **Operator** for Pro subscribers – an experimental AI agent capable of browsing websites and performing actions. (*special*) 193 | - Google introduced **Gemini Flash Thinking 0121**, an enhanced reasoning model that secured the top spot in the Arena Chatbots rankings. 194 | - DeepSeek open-sourced the reasoning models **R1** and **R1-Zero**, which demonstrated capabilities similar to **o1** across various domains at a fraction of the cost. Additionally, smaller distilled models were released, achieving high performance relative to their size. (*special*) 195 | - Google published a research paper on a new language model architecture called **Titans**, designed to enable models to retain both short- and long-term memory. This architecture significantly improves processing for extended context windows. (*special*) 196 | - DeepSeek open-sourced a fully multimodal model, **Janus Pro 7B**, which supports both text and image generation. 197 | - Alibaba unveiled **Qwen2.5-Max**, a large language model that surpasses several leading models, including **DeepSeek-V3**, **GPT-4o**, and **Claude 3.5**. Additionally, the **Qwen2.5-1M** series was open-sourced, capable of processing up to one million tokens, along with the **Qwen2.5-VL** vision model series in three different sizes. 198 | - OpenAI made the **o3 mini** reasoning model available to all users, including the free tier, featuring three reasoning levels. The model matches or comes close to o1 in several benchmarks, significantly surpasses it in coding, and remains significantly faster and more cost-efficient. (*special*) 199 | 200 | ## February 2025 201 | - xAI launches **Grok 3**, **Grok 3 Reasoning** and **Grok 3 mini**, next-generation AI models trained with 10 times the computing power of Grok 2, significantly improving SOTA performance. They include "Think" and "Big Brain" modes for advanced reasoning, as well as **DeepSearch** for autonomous web searches. (*special*) 202 | - Anthropic introduces **Claude 3.7** and **Claude 3.7 Thinking**, a new model with enhanced coding performance, support for "Extended Thinking" mode, and the ability to analyze reasoning processes. (*special*) 203 | - OpenAI unveils **Deep Research**, a tool for autonomous research, enabling real-time web searches and comprehensive report generation. (*special*) 204 | - Google releases **Gemini 2.0 Flash**, **Gemini 2.0 Flash-Lite Preview**, and **Gemini 2.0 Pro Experimental**. 205 | - Alibaba launches **QwQ-Max** – a reasoning model based on Qwen2.5-Max, offering improved analytical and logical capabilities. 206 | - Microsoft presents **Phi4-mini** and **Phi4 Multimodal**, lightweight models (3.8B and 5.6B) with enhanced performance, including support for multimodal inputs. 207 | - OpenAI releases **GPT-4.5**, featuring advanced pattern recognition and significantly reduced hallucinations, improving accuracy and reliability. (*special*) 208 | 209 | ## March 2025 210 | - Google introduced **Gemini 2.5 Pro**, an experimental "Thinking model" with advanced reasoning and planning capabilities, a 1 million token context window, achieving top rankings across several key benchmarks. (*special*) 211 | - Google launched the Gemma 3 series, featuring open-source multimodal models in various parameter sizes, a 128K context window, multi-language support, and integrated image and video understanding capabilities. 212 | - OpenAI integrated **GPT-4o Image Generation**, enabling high-fidelity text-to-image creation, text rendering within images, and more. (*special*) 213 | - Google expanded experimental image generation and editing within **Gemini 2.0 Flash Experimental**, enabling image generation and editing, including enhanced text creation capabilities. (*special*) 214 | - Alibaba released **QwQ-32B**, an open-source 32B parameter reasoning model with exceptional math and coding performance, rivaling much larger models. 215 | - Alibaba released the **Qwen2.5-VL 32B**, open-source vision-language model with robust capabilities in visual analysis, text-in-image understanding, and visual agent tasks. 216 | - DeepSeek updated its open-source MoE model with **DeepSeek-V3-0324**, featuring enhanced reasoning, coding, and math capabilities, positioning it as a top-tier base model. 217 | - Sesame AI unveiled its **Conversational Speech Model (CSM)**, enabling remarkably human-like, real-time voice interaction, incorporating emotional nuances, natural pauses, laughter, and contextual memory. (*special*) 218 | 219 | ## April 2025 220 | - Meta releases **Llama 4** in three sizes with a context window of 10 million tokens and medium performance. 221 | - Google launches **Gemini 2.5 Flash**, with a dynamic reasoning mode that allows tuning the reasoning level or disabling it as needed. 222 | - Amazon introduces **Nova Act**, a new framework for building multi-step autonomous agents. 223 | - OpenAI releases **GPT-4.1** in three sizes, with a context window of 1 million tokens. 224 | - OpenAI introduces **O3 full** and **O4 mini**, highly advanced models for reasoning, math, and coding. 225 | - Midjourney launches **v7**, with higher image quality and more precise control over style. 226 | - A series of video model updates - **Veo 2.0** (Google), **Runway Gen-4**, **Vidu Q1**, and **Kling 2.0** – a leap forward in high-quality video generation, with improvements in response times, realism, and style. 227 | - Alibaba releases **Qwen 3** as open source, in various sizes, with very impressive capabilities for their size. (*special*) 228 | 229 | ## May 2025 230 | - Microsoft launches the **Phi-4 reasoning** series as open source, small yet high-quality models that incorporate reasoning. 231 | - Suno releases **Suno 4.5**, fixing shimmer noise and improving audio decay stability in long tracks. 232 | - Anthropic releases **Claude 4 Opus** and **Claude Sonnet 4**: Opus 4 offers a Hybrid "Deep Thought" mode with enhanced long-term context and 7-hour autonomous operation; Sonnet 4 focuses on improved math and coding performance. (*special*) 233 | - Google releases **Veo 3**, a video generation model for synchronized 4K video with natural audio integration, and **Imagen 4**, an advanced image model with deeper contextual understanding and artistic style support. (*special*) 234 | - OpenAI releases **Codex**, an autonomous code agent in ChatGPT, powered by the o3 model, for writing code, debugging, testing, and creating GitHub Pull Requests. 235 | - Google releases **Jules**, an asynchronous autonomous coding agent on Gemini 2.5 Pro, analyzing repositories and creating GitHub Pull Requests. 236 | - Google releases **Gemini 2.5 Pro** (Deep Think Mode) and **Gemini 2.5 Flash**, featuring improved reasoning, native audio support, extended context, and high-frequency task handling. 237 | - OpenAI updates **Operator** to use the **o3** model, achieving SOTA on OSWorld benchmarks and enhancing autonomous browser capabilities. 238 | - DeepSeek open-sources **R1-0528**, a code-and-inference model with near–o4-mini performance and moderate computational needs. 239 | - Google DeepMind launches **AlphaEvolve**, an autonomous code-optimizer using evolutionary strategies with LLMs, achieving SOTA on 75% of math problems and discovering enhanced algorithms 20% of the time. (*special*) 240 | - Google releases **Gemini Diffusion**, an experimental text diffusion model achieving high-speed text generation with enhanced control and creativity via noise refinement. (*special*) 241 | - Google introduces **Gemma 3n**, an open-source generative AI model for on-device use, with an efficient architecture and multi-modal (audio, text, visual) capabilities. 242 | 243 | 244 | ## June 2025 245 | - Google releases **Gemini 2.5 Pro** (final production-ready version), which leads benchmarks across the board. 246 | - ElevenLabs rolls out **Eleven v3 (alpha)** TTS with fine grained emotion control and support for 70+ languages. 247 | - OpenAI debuts **o3 pro**, an enhanced reasoning model offering extended context and real-time tool integrations. 248 | 249 | ## July 2025 250 | - xAI releases **Grok 4**, achieving a new SOTA of 15.9% on ARC-AGI v2 and 25.4% on Humanity’s Last Exam. (*special*) 251 | - OpenAI unveils the **ChatGPT Agent**, embedding autonomous coding, web research and tool use directly within the chat interface. (*special*) 252 | - An experimental OpenAI model secures a **gold medal** at IMO 2025 without any external tools. (*special*) 253 | - Google introduces **Gemini Deep Think**, which also earns an IMO 2025 gold by solving five of six problems with parallel reasoning. (*special*) 254 | - Alibaba open-sources two variants, **Qwen3-235B-A22B-Instruct-2507** (instruction-tuned) and **Qwen3-Coder**, for general LLM use and automated code generation. 255 | - Moonshot AI debuts **Kimi K2**, a Chinese LLM praised for its open-research focus and robust performance. 256 | - Chinese startup Zhipu open-sources **GLM-4.5**, a 130 B-parameter model tailored for intelligent-agent applications. 257 | 258 | ## August 2025 259 | - Google introduced **Gemini 2.5 Deep Think**, a special "extended thinking" mode for solving complex problems and exploring alternatives. (*special*) 260 | - Anthropic released **Claude Opus 4.1**, an upgrade focused on improving agentic capabilities and real-world coding. 261 | - Google DeepMind announced **Genie 3.0**, a "world model" for creating interactive 3D environments from text, maintaining consistency for several minutes. (*special*) 262 | - OpenAI released **gpt-oss-120b** and **gpt-oss-20b**, a family of open-source models with high reasoning capabilities, optimized to run on accessible hardware. 263 | - OpenAI launched **GPT-5**, the company's next-generation model, with significant improvements in coding and a dynamic "thinking" mode to reduce hallucinations. 264 | - DeepSeek released **DeepSeek V3.1**, a hybrid model combining fast and slow "thinking" modes to improve performance in agentic tasks and tool use. 265 | - Google launched a preview of **Gemini 2.5 Flash Image** (showcased as *nano-banana*), an advanced model for precise image editing, merging, and maintaining character consistency. (*special*) 266 | 267 | 268 | ## September 2025 269 | 270 | - ByteDance released **Seedream 4.0**, a next-generation image model unifying high-quality text-to-image generation and natural-language image editing. 271 | - An advanced Gemini variant, reported as **Gemini 2.5 - Deep Think**, achieved gold-medal-level performance at the ICPC World Finals programming contest. (*special*) 272 | - OpenAI reported a reasoning and code model achieved a perfect score (12/12) in ICPC testing. (*special*) 273 | - Suno released **Suno v5**, an upgrade in music generation with studio-grade fidelity and more natural-sounding vocals. 274 | - Alibaba unveiled **Qwen-3-Max**, its flagship model with over a trillion parameters, focusing on long context and agent capabilities. 275 | - **Wan 2.5** was released, a generative video model focused on multi-shot consistency and character animation. 276 | - Anthropic announced **Claude Sonnet 4.5**, a model optimized for coding, agent construction, and improved reasoning. 277 | - OpenAI released **Sora 2**, a flagship video and audio generation model with improved physical modeling and synchronized sound. 278 | - DeepSeek released **DeepSeek-V3.2-Exp** 279 | - OpenAI and NVIDIA announced a strategic partnership for NVIDIA to supply at least **10 gigawatts** of AI systems for OpenAI's infrastructure. (*special*) 280 | 281 | ## October 2025 282 | 283 | - Figure unveiled **Figure 03**, a humanoid robot designed for domestic and general-purpose tasks. 284 | - Google released a **Gemini model for computer control**, achieving state-of-the-art (SOTA) performance in GUI automation. 285 | - Anthropic released **Claude 4.5 Haiku**, a fast, cost-effective model for high-volume, low-latency applications. 286 | - OpenAI announced **ChatGPT Atlas**, an AI-native web browser with a built-in "Agent Mode" for task automation. 287 | - 1X announced **Neo**, a humanoid robot marketed as the first consumer-ready model for home use. (*special*) 288 | 289 | 290 | ## November 2025 291 | 292 | - Moonshot AI released **Kimi K2 Thinking**, an open model setting new records in reasoning benchmarks. 293 | - OpenAI launched **GPT 5.1**, featuring specialized "Thinking" and "Instant" modes with expanded context. 294 | - xAI released **Grok 4.1**, combining high EQ with strong logic to top the LM Arena leaderboard. 295 | - Google debuted **Gemini 3.0**, a flagship "thinking" model that claimed the top spot on major benchmarks. 296 | - OpenAI introduced **GPT 5.1 Codex Max**, an agentic model built specifically for long-term coding tasks. 297 | - Google released **Nano Banana Pro**, a superior image generation and editing model based on Gemini 3. (*special*) 298 | - Anthropic announced **Claude Opus 4.5**, delivering elite coding and agentic performance at a significantly reduced price. (*special*) 299 | - Black Forest Labs launched **FLUX 2**, a high-performance open-weight image generation model. 300 | - DeepSeek released **DeepSeekMath-V2** as open source, achieving gold-medal performance in math olympiads. (*special*) 301 | - Microsoft open-sourced **Fara-7B**, a small model optimized for browser agents and computer control. 302 | - **Poetiq** shatters the **ARC-AGI-2** benchmark with a score of over 60%, surpassing the human average. 303 | -------------------------------------------------------------------------------- /_data/timeline.yml: -------------------------------------------------------------------------------- 1 | - year: 2022 2 | events: 3 | - date: February 4 | info: 5 | - text: Midjourney v1 6 | - date: March 7 | info: 8 | - text: OpenAI releases text-davinci-002 and code-davinci-002 with an API approach. 9 | - date: April 10 | info: 11 | - text: Midjourney v2 12 | - text: DALL-E 2 is announced for gradual release. 13 | special: true 14 | - date: July 15 | info: 16 | - text: Midjourney v3 is launched. 17 | - date: August 18 | info: 19 | - text: Stable Diffusion 1.4 is released. 20 | - date: October 21 | info: 22 | - text: Stable Diffusion 1.5 becomes available. 23 | special: true 24 | - date: November 25 | info: 26 | - text: ChatGPT, a chatbot by OpenAI using GPT-3.5, is released to the public and quickly becomes a viral sensation. 27 | special: true 28 | - text: Midjourney v4 is released. 29 | - text: Stable Diffusion 2.0 is launched. 30 | - date: December 31 | info: 32 | - text: Stable Diffusion 2.1 is released. 33 | - year: 2023 34 | events: 35 | - date: February 36 | info: 37 | - text: Meta releases the LLaMA language model as open-source for research purposes. The model is later leaked. 38 | special: true 39 | - text: Microsoft gradually releases Bing AI, an AI chat based on an upgraded GPT model integrating internet search. 40 | - date: March 41 | info: 42 | - text: Midjourney v5 is launched. 43 | - text: OpenAI's GPT-4 model is partially released, featuring multimodal image analysis and improved multi-language support. 44 | special: true 45 | - text: Google releases the AI chat Bard in a limited capacity, based on the LaMDA language model. 46 | - date: April 47 | info: 48 | - text: Adobe releases the Firefly image creation model as a beta version to a waiting list. The model allowed a variety of capabilities including text formatting. 49 | - date: May 50 | info: 51 | - text: Midjourney v5.1 is released. 52 | - text: Google announces an upgrade to Bard, moving it to the upgraded PaLM 2 language model. It will support 180 countries and many languages. 53 | - date: June 54 | info: 55 | - text: Midjourney v5.2 is launched. 56 | - date: July 57 | info: 58 | - text: Stable Diffusion XL 1.0 is released. 59 | - text: Anthropic announces a new version of their large language model - Claude 2. 60 | - text: Meta releases the LLaMA 2 open source language model to the general public in a variety of sizes. 61 | - date: October 62 | info: 63 | - text: DALL-E 3 is released. 64 | - text: Adobe releases Firefly 2. 65 | - date: November 66 | info: 67 | - text: Stable Diffusion XL Turbo is released - A fast model that allows the creation of an image in one step in real-time. 68 | - date: December 69 | info: 70 | - text: Midjourney v6 is launched. 71 | - text: Google upgrades Bard in limited areas, moving it to be based on the upgraded Gemini Pro language model. 72 | - text: X Corporation launches Grok AI chatbot for paid subscribers in English language. 73 | - year: 2024 74 | events: 75 | - date: February 76 | info: 77 | - text: Stability AI announces Stable Diffusion 3 (gradually released to waiting list). 78 | - text: Google upgrades the artificial intelligence chat in Bard, basing it on the new Gemini Pro model, in all available languages. Google replaces "Bard" with "Gemini". 79 | - text: Google announces the Gemini Pro 1.5 multimodal language model capable of parsing up to a million tokens, as well as parsing video and images. The model is gradually released to developers on a waiting list. 80 | special: true 81 | - text: OpenAI announces the Sora model that produces videos up to a minute long. The model is not released to the public at this time. 82 | special: true 83 | - date: March 84 | info: 85 | - text: X Corporation announces the upcoming release of the Grok 1.5 open source model. 86 | - text: Anthropic announces Claude 3, a new version of their large language model. The version is deployed in 3 different sizes, with the largest model performing better than GPT-4. 87 | - text: Suno AI, which develops a model for creating music, releases Suno v3 to the general public. 88 | - date: April 89 | info: 90 | - text: Stability AI releases a new update to the music creation model - Stable Audio 2.0. 91 | - text: X Corporation releases an upgrade to its language model, Grok-1.5V, which integrates high-level image recognition. In the test presented by the company, the model is the best in identifying and analyzing images compared to other models. 92 | - text: The Mistral company releases its new model Mixtral 8x22B as open source. This is the most powerful model among the open source models and it contains 141 billion parameters but uses a method that allows more economical use. 93 | - text: Meta releases the LLaMA 3 model as open source in sizes 8B and 70B parameters. The large model shows better performance than Claude 3 Sonnet and Gemini Pro 1.5 in several measures. Meta is expected to later release larger models with 400 billion parameters and more. 94 | - text: Microsoft releases the Phi-3-mini model in open source. The model comes in a reduced version of 3.8B parameters, which allows it to run on mobile devices as well, and it presents capabilities similar to GPT-3.5. 95 | special: true 96 | - text: Adobe announces its new image creation model Firefly 3. 97 | - text: The startup Reka AI presents a series of multimodal language models in 3 sizes. The models are capable of processing video, audio and images. The large model featured similar capabilities to GPT-4. 98 | - text: Apple releases as full open source a series of small language models under the name OpenELM. The models are available in four weights between 270 million and 3 billion parameters. 99 | - date: May 100 | info: 101 | - text: OpenAI announces the GPT-4o model that presents full multimodal capabilities, including receiving and creating text, images, and audio. The model presents an impressive ability to speak with a high response speed and in natural language. The model is 2 times more efficient than the GPT-4 Turbo model, and has better capabilities for languages other than English. 102 | special: true 103 | - text: 'Google announces a large number of AI features in its products. The main ones: increasing the token limit to 2 million for Gemini 1.5 to waiting list, releasing a smaller and faster Gemini Flash 1.5 model. Revealing the latest image creation model Imagen 3, music creation model Music AI and video creation model Veo. And the announcement of the Astra model with multimodal capabilities for realtime audio and video reception.' 104 | - text: 'Microsoft announces Copilot+ for dedicated computers, which will allow a full search of the user''s history through screenshots of the user''s activity. The company also released as open source the SLMs that display impressive capabilities in a minimal size: Phi-3 Small, Phi-3 Medium, and Phi-3 Vision which includes image recognition capability.' 105 | - text: Meta introduces Chameleon, a new multimodal model that seamlessly renders text and images. 106 | - text: Mistral AI releases a new open source version of its language model Mistral-7B-Instruct-v0.3. 107 | - text: Google announces AI Overviews intended to give a summary of the relevant information in Google search. 108 | special: true 109 | - text: Suno AI releases an updated music creation model Suno v3.5. 110 | - text: Mistral AI releases a new language model designed for coding Codestral in size 22B. 111 | - date: June 112 | info: 113 | - text: Stability AI releases its updated image creation model Stable Diffusion 3 in a medium version in size 2B parameters. 114 | - text: Apple announces Apple Intelligence, an AI system that will be integrated into the company's devices and will combine AI models of different sizes for different tasks. 115 | - text: DeepSeekAI publishes the DeepSeekCoderV2 open source language model which presents similar coding capabilities to models such as GPT-4, Claude 3 Opus and more. 116 | - text: Runway introduces Gen3 Alpha, a new AI model for video generation. 117 | - text: Anthropic releases the Claude Sonnet 3.5 model, which presents better capabilities than other models with low resource usage. 118 | special: true 119 | - text: Microsoft releases in open source a series of image recognition models called Florence 2. 120 | - text: Google announces Gemma 2 open source language models with 9B and 27B parameter sizes. Also, the company opens the context window capabilities to developers for up to 2 million tokens. 121 | - date: July 122 | info: 123 | - text: OpenAI has released a miniaturized model called GPT-4o mini that presents high capabilities at a low cost 124 | - text: Meta releases as open source the llama 3.1 model in sizes 8B, 70B and 405B. The large model features the same capabilities as the best closed source models 125 | special: true 126 | - text: 'mistral ai releases three new models: Codestral Mamba, Mistral NeMo and Mathstral designed for mathematics' 127 | - text: Google DeepMind has unveiled two new AI systems that won silver medals at this year's International Mathematical Olympiad (IMO), AlphaProof and AlphaGeometry 2. 128 | special: true 129 | - text: OpenAI launched SearchGPT, an integrated web search 130 | - text: Startup Udio has released Udio v1.5, an updated version of its music creation model 131 | - text: Mistral AI has released a large language model Mistral Large 2 in size 123B, which presents capabilities close to the closed SOTA models. 132 | special: true 133 | - text: Midjourney v6.1 is released 134 | - text: Google releases the Gemma 2 2B model as open source. The model demonstrates better capabilities than much larger models. 135 | - date: August 136 | info: 137 | - text: '"Black Forest Labs" releases weights for an image creation model named Flux, which shows better performance than similar closedsource models.' 138 | - text: OpenAI released a new version of its model, GPT-4o 0806, achieving 100% success in generating valid JSON output. 139 | - text: Google's image generation model, Imagen 3, has been released. 140 | - text: xAI Corporation has launched the models Grok 2 and Grok 2 mini, which demonstrate performance on par with leading SOTA models in the market. 141 | - text: Microsoft has introduced its small language models, Phi 3.5, in three versions, each showcasing impressive performance relative to their size. 142 | - text: 'Google has introduced three new experimental AI models: Gemini 1.5 Flash8B, Gemini 1.5 Pro Enhanced, and Gemini 1.5 Flash Updated.' 143 | - text: Ideogram 2.0 has been released, offering image generation capabilities that surpass those of other leading models. 144 | - text: Luma has unveiled the Dream Machine 1.5 model for video creation. 145 | - date: September 146 | info: 147 | - text: The French AI company Mistral has introduced Pixtral12B, its first multimodal model capable of processing both images and text. 148 | - text: 'OPENAI has released two nextgeneration AI models to its subscribers: o1 preview and o1 mini. These models show a significant improvement in performance, particularly in tasks requiring reasoning, including coding, mathematics, GPQA, and more.' 149 | special: true 150 | - text: Chinese company Alibaba releases the Qwen 2.5 model in various sizes, ranging from 0.5B to 72B. The models demonstrate capabilities comparable to much larger models. 151 | - text: The video generation model KLING 1.5 has been released. 152 | - text: OpenAI launches the advanced voice mode of GPT4o for all subscribers. 153 | - text: Meta releases Llama 3.2 in sizes 1B, 3B, 11B and 90B, featuring image recognition capabilities for the first time. 154 | - text: Google has rolled out new model updates ready for deployment, Gemini Pro 1.5 002 and Gemini Flash 1.5 002, showcasing significantly improved longcontext processing. 155 | - text: Kyutai releases two opensource versions of its voicetovoice model, Moshi. 156 | - text: Google releases an update to its AI tool NotebookLM that enables users to create podcasts based on their own content. 157 | - text: Mistral AI launches a 22B model named Mistral Small. 158 | - date: October 159 | info: 160 | - text: Flux 1.1 Pro is released, showcasing advanced capabilities for image creation. 161 | - text: Meta unveils Movie Gen, a new AI model that generates videos, images, and audio from text input. 162 | - text: Pika introduces Video Model 1.5 along with "Pika Effects." 163 | - text: Adobe announces its video creation model, Firefly Video. 164 | - text: Startup Rhymes AI releases Aria, an opensource, multimodal model exhibiting capabilities similar to comparably sized proprietary models. 165 | - text: Meta releases an opensource speechtospeech language model named Meta Spirit LM. 166 | - text: Mistral AI introduces Ministral, a new model available in 3B and 8B parameter sizes. 167 | - text: Janus AI, a multimodal language model capable of recognizing and generating both text and images, is released as open source by DeepSeekAI. 168 | - text: Google DeepMind and MIT unveil Fluid, a texttoimage generation model with industryleading performance at a scale of 10.5B parameters. 169 | - text: Stable Diffusion 3.5 is released in three sizes as open source. 170 | - text: Anthropic launches Claude 3.5 Sonnet New, demonstrating significant advancements in specific areas over its previous version, and announces Claude 3.5 Haiku. 171 | - text: Anthropic announces an experimental feature for computer use with a public beta API. 172 | - text: The texttoimage model Recraft v3 has been released to the public, ranking first in benchmarks compared to similar models. 173 | - text: OpenAI has launched Search GPT, allowing users to perform web searches directly within the platform. 174 | - date: November 175 | info: 176 | - text: Alibaba released its new model, QwQ 32B Preview, which integrates reasoning capabilities before responding. The model competes with, and sometimes surpasses, OpenAI's o1-preview model. 177 | - text: Alibaba opensourced the model Qwen2.5 Coder 32B, which offers comparable capabilities to leading proprietary language models in the coding domain. 178 | - text: DeepSeek unveiled its new AI model, DeepSeek-R1-Lite-Preview, which incorporates reasoning capabilities and delivers impressive performance on the AIME and MATH benchmarks, matching the level of OpenAI's o1-preview. 179 | - text: Suno upgraded its AIpowered music generator to v4, introducing new features and performance improvements. 180 | - text: Mistral AI launched the Pixtral Large model, a multimodal language model excelling in image recognition and advanced performance metrics, and an update to Mistral Large, 2411. 181 | - text: Google introduced two experimental models, gemini-exp-1114 and gemini-exp-1121, currently leading the arena chatbot with enhanced performance. 182 | - text: Anthropic launches Claude 3.5 Haiku and Visual PDF Analysis in Claude. 183 | - date: December 184 | info: 185 | - text: Amazon introduced a new series of models called NOVA, designed for text, image, and video processing. 186 | - text: OpenAI released SORA, a video generation model, along with the full version of O1 and O1 Pro for advanced subscribers. Additionally, the company launched a live video mode for GPT4o. 187 | special: true 188 | - text: Google unveiled the experimental model Gemini-Exp-1206, which ranked first in the chatbot leaderboard. 189 | - text: Google launched Gemini 2.0 Flash in beta. This model leads benchmarks and outperforms the previous version, Gemini Pro 1.5. Additionally, Google introduced live speech and video mode and announced built-in image generation capabilities within the model. 190 | special: true 191 | - text: Google revealed Gemini-2.0-Flash-Thinking, a thinking model based on Gemini 2.0 Flash, which secured second place in the chatbot leaderboard. 192 | special: true 193 | - text: Google introduced Veo 2, a beta version video generation model capable of producing 4K videos up to two minutes long. The model outperformed SORA in human evaluations. Additionally, Google updated Imagen 3, offering enhanced image quality and realism. 194 | special: true 195 | - text: xAI integrated Aurora, a new model for generating high-quality and realistic images. 196 | - text: Microsoft open-sourced the Phi4 model, sized at 14B, showcasing impressive capabilities for its size. 197 | - text: Meta released Llama 3.3 70B, a model offering performance comparable to Llama 3.1 405B. 198 | - text: Google launched a multi-modal open-source model called PaliGemma 2, integrated with existing Gemma models. 199 | - text: Pika Labs released 2.0, the latest version of its AI-powered video generator. 200 | - text: Meta introduced Apollo, a video generation model available in three different sizes. 201 | - text: Deepseek open-sourced Deepseek V3, a model with 671B parameters that surpasses closed-source SOTA models across several benchmarks. 202 | special: true 203 | - text: Alibaba unveiled QVQ-72B-Preview, a cutting-edge thinking model capable of analyzing images, featuring SOTA-level performance. 204 | special: true 205 | - text: OpenAI announced O3, a groundbreaking AI model achieving 87.5% in the ARC-AGI benchmark, 25.2% in the Frontier Math Benchmark (compared to under 2% in previous models), and 87.7% in Ph.D.-level science questions. A cost-effective version, O3 Mini, is expected in January 2025, with performance similar to O1, alongside improved speed and efficiency. 206 | special: true 207 | - text: The video generation model Kling 1.6 was released, offering significant performance enhancements. 208 | - year: 2025 209 | events: 210 | - date: January 211 | info: 212 | - text: OpenAI released Operator for Pro subscribers – an experimental AI agent capable of browsing websites and performing actions. 213 | special: true 214 | - text: Google introduced Gemini Flash Thinking 0121, an enhanced reasoning model that secured the top spot in the Arena Chatbots rankings. 215 | - text: DeepSeek open-sourced the reasoning models R1 and R1-Zero, which demonstrated capabilities similar to o1 across various domains at a fraction of the cost. Additionally, smaller distilled models were released, achieving high performance relative to their size. 216 | special: true 217 | - text: Google published a research paper on a new language model architecture called Titans, designed to enable models to retain both short- and long-term memory. This architecture significantly improves processing for extended context windows. 218 | special: true 219 | - text: DeepSeek open-sourced a fully multimodal model, Janus Pro 7B, which supports both text and image generation. 220 | - text: Alibaba unveiled Qwen2.5-Max, a large language model that surpasses several leading models, including DeepSeek-V3, GPT-4o, and Claude 3.5. Additionally, the Qwen2.5-1M series was open-sourced, capable of processing up to one million tokens, along with the Qwen2.5-VL vision model series in three different sizes. 221 | - text: OpenAI made the o3 mini reasoning model available to all users, including the free tier, featuring three reasoning levels. The model matches or comes close to o1 in several benchmarks, significantly surpasses it in coding, and remains significantly faster and more cost-efficient. 222 | special: true 223 | - date: February 2025 224 | info: 225 | - text: xAI launches Grok 3, Grok 3 Reasoning and Grok 3 mini, next-generation AI models trained with 10 times the computing power of Grok 2, significantly improving SOTA performance. They include "Think" and "Big Brain" modes for advanced reasoning, as well as DeepSearch for autonomous web searches. 226 | special: true 227 | - text: Anthropic introduces Claude 3.7 and Claude 3.7 Thinking, a new model with enhanced coding performance, support for "Extended Thinking" mode, and the ability to analyze reasoning processes. 228 | special: true 229 | - text: OpenAI unveils Deep Research, a tool for autonomous research, enabling real-time web searches and comprehensive report generation. 230 | special: true 231 | - text: Google releases Gemini 2.0 Flash, Gemini 2.0 Flash-Lite Preview, and Gemini 2.0 Pro Experimental. 232 | - text: Alibaba launches QwQ-Max – a reasoning model based on Qwen2.5-Max, offering improved analytical and logical capabilities. 233 | - text: Microsoft presents Phi4-mini and Phi4 Multimodal, lightweight models (3.8B and 5.6B) with enhanced performance, including support for multimodal inputs. 234 | - text: OpenAI releases GPT-4.5, featuring advanced pattern recognition and significantly reduced hallucinations, improving accuracy and reliability. 235 | special: true 236 | - date: March 2025 237 | info: 238 | - text: Google introduced Gemini 2.5 Pro, an experimental "Thinking model" with advanced reasoning and planning capabilities, a 1 million token context window, achieving top rankings across several key benchmarks. 239 | special: true 240 | - text: Google launched the Gemma 3 series, featuring open-source multimodal models in various parameter sizes, a 128K context window, multi-language support, and integrated image and video understanding capabilities. 241 | - text: OpenAI integrated GPT-4o Image Generation, enabling high-fidelity text-to-image creation, text rendering within images, and more. 242 | special: true 243 | - text: Google expanded experimental image generation and editing within Gemini 2.0 Flash Experimental, enabling image generation and editing, including enhanced text creation capabilities. 244 | special: true 245 | - text: Alibaba released QwQ-32B, an open-source 32B parameter reasoning model with exceptional math and coding performance, rivaling much larger models. 246 | - text: Alibaba released the Qwen2.5-VL 32B, open-source vision-language model with robust capabilities in visual analysis, text-in-image understanding, and visual agent tasks. 247 | - text: DeepSeek updated its open-source MoE model with DeepSeek-V3-0324, featuring enhanced reasoning, coding, and math capabilities, positioning it as a top-tier base model. 248 | - text: Sesame AI unveiled its Conversational Speech Model (CSM), enabling remarkably human-like, real-time voice interaction, incorporating emotional nuances, natural pauses, laughter, and contextual memory. 249 | special: true 250 | - date: April 2025 251 | info: 252 | - text: Meta releases Llama 4 in three sizes with a context window of 10 million tokens and medium performance. 253 | - text: Google launches Gemini 2.5 Flash, with a dynamic reasoning mode that allows tuning the reasoning level or disabling it as needed. 254 | - text: Amazon introduces Nova Act, a new framework for building multi-step autonomous agents. 255 | - text: OpenAI releases GPT-4.1 in three sizes, with a context window of 1 million tokens. 256 | - text: OpenAI introduces O3 full and O4 mini, highly advanced models for reasoning, math, and coding. 257 | - text: Midjourney launches v7, with higher image quality and more precise control over style. 258 | - text: A series of video model updates - Veo 2.0 (Google), Runway Gen-4, Vidu Q1, and Kling 2.0 – a leap forward in high-quality video generation, with improvements in response times, realism, and style. 259 | - text: Alibaba releases Qwen 3 as open source, in various sizes, with very impressive capabilities for their size. 260 | special: true 261 | - date: May 2025 262 | info: 263 | - text: Microsoft launches the Phi-4 reasoning series as open source, small yet high-quality models that incorporate reasoning. 264 | - text: Suno releases Suno 4.5, fixing shimmer noise and improving audio decay stability in long tracks. 265 | - text: 'Anthropic releases Claude 4 Opus and Claude Sonnet 4: Opus 4 offers a Hybrid "Deep Thought" mode with enhanced long-term context and 7-hour autonomous operation; Sonnet 4 focuses on improved math and coding performance.' 266 | special: true 267 | - text: Google releases Veo 3, a video generation model for synchronized 4K video with natural audio integration, and Imagen 4, an advanced image model with deeper contextual understanding and artistic style support. 268 | special: true 269 | - text: OpenAI releases Codex, an autonomous code agent in ChatGPT, powered by the o3 model, for writing code, debugging, testing, and creating GitHub Pull Requests. 270 | - text: Google releases Jules, an asynchronous autonomous coding agent on Gemini 2.5 Pro, analyzing repositories and creating GitHub Pull Requests. 271 | - text: Google releases Gemini 2.5 Pro (Deep Think Mode) and Gemini 2.5 Flash, featuring improved reasoning, native audio support, extended context, and high-frequency task handling. 272 | - text: OpenAI updates Operator to use the o3 model, achieving SOTA on OSWorld benchmarks and enhancing autonomous browser capabilities. 273 | - text: DeepSeek open-sources R1-0528, a code-and-inference model with near–o4-mini performance and moderate computational needs. 274 | - text: Google DeepMind launches AlphaEvolve, an autonomous code-optimizer using evolutionary strategies with LLMs, achieving SOTA on 75% of math problems and discovering enhanced algorithms 20% of the time. 275 | special: true 276 | - text: Google releases Gemini Diffusion, an experimental text diffusion model achieving high-speed text generation with enhanced control and creativity via noise refinement. 277 | special: true 278 | - text: Google introduces Gemma 3n, an open-source generative AI model for on-device use, with an efficient architecture and multi-modal (audio, text, visual) capabilities. 279 | - date: June 2025 280 | info: 281 | - text: Google releases Gemini 2.5 Pro (final production-ready version), which leads benchmarks across the board. 282 | - text: ElevenLabs rolls out Eleven v3 (alpha) TTS with fine grained emotion control and support for 70+ languages. 283 | - text: OpenAI debuts o3 pro, an enhanced reasoning model offering extended context and real-time tool integrations. 284 | - date: July 2025 285 | info: 286 | - text: xAI releases Grok 4, achieving a new SOTA of 15.9% on ARC-AGI v2 and 25.4% on Humanity’s Last Exam. 287 | special: true 288 | - text: OpenAI unveils the ChatGPT Agent, embedding autonomous coding, web research and tool use directly within the chat interface. 289 | special: true 290 | - text: An experimental OpenAI model secures a gold medal at IMO 2025 without any external tools. 291 | special: true 292 | - text: Google introduces Gemini Deep Think, which also earns an IMO 2025 gold by solving five of six problems with parallel reasoning. 293 | special: true 294 | - text: Alibaba open-sources two variants, Qwen3-235B-A22B-Instruct-2507 (instruction-tuned) and Qwen3-Coder, for general LLM use and automated code generation. 295 | - text: Moonshot AI debuts Kimi K2, a Chinese LLM praised for its open-research focus and robust performance. 296 | - text: Chinese startup Zhipu open-sources GLM-4.5, a 130 B-parameter model tailored for intelligent-agent applications. 297 | - date: August 2025 298 | info: 299 | - text: Google introduced Gemini 2.5 Deep Think, a special "extended thinking" mode for solving complex problems and exploring alternatives. 300 | special: true 301 | - text: Anthropic released Claude Opus 4.1, an upgrade focused on improving agentic capabilities and real-world coding. 302 | - text: Google DeepMind announced Genie 3.0, a "world model" for creating interactive 3D environments from text, maintaining consistency for several minutes. 303 | special: true 304 | - text: OpenAI released gpt-oss-120b and gpt-oss-20b, a family of open-source models with high reasoning capabilities, optimized to run on accessible hardware. 305 | - text: OpenAI launched GPT-5, the company's next-generation model, with significant improvements in coding and a dynamic "thinking" mode to reduce hallucinations. 306 | - text: DeepSeek released DeepSeek V3.1, a hybrid model combining fast and slow "thinking" modes to improve performance in agentic tasks and tool use. 307 | - text: Google launched a preview of Gemini 2.5 Flash Image (showcased as *nano-banana*), an advanced model for precise image editing, merging, and maintaining character consistency. 308 | special: true 309 | - date: September 2025 310 | info: 311 | - text: ByteDance released Seedream 4.0, a next-generation image model unifying high-quality text-to-image generation and natural-language image editing. 312 | - text: An advanced Gemini variant, reported as Gemini 2.5 - Deep Think, achieved gold-medal-level performance at the ICPC World Finals programming contest. 313 | special: true 314 | - text: OpenAI reported a reasoning and code model achieved a perfect score (12/12) in ICPC testing. 315 | special: true 316 | - text: Suno released Suno v5, an upgrade in music generation with studio-grade fidelity and more natural-sounding vocals. 317 | - text: Alibaba unveiled Qwen-3-Max, its flagship model with over a trillion parameters, focusing on long context and agent capabilities. 318 | - text: Wan 2.5 was released, a generative video model focused on multi-shot consistency and character animation. 319 | - text: Anthropic announced Claude Sonnet 4.5, a model optimized for coding, agent construction, and improved reasoning. 320 | - text: OpenAI released Sora 2, a flagship video and audio generation model with improved physical modeling and synchronized sound. 321 | - text: DeepSeek released DeepSeek-V3.2-Exp 322 | - text: OpenAI and NVIDIA announced a strategic partnership for NVIDIA to supply at least 10 gigawatts of AI systems for OpenAI's infrastructure. 323 | special: true 324 | - date: October 2025 325 | info: 326 | - text: Figure unveiled Figure 03, a humanoid robot designed for domestic and general-purpose tasks. 327 | - text: Google released a Gemini model for computer control, achieving state-of-the-art (SOTA) performance in GUI automation. 328 | - text: Anthropic released Claude 4.5 Haiku, a fast, cost-effective model for high-volume, low-latency applications. 329 | - text: OpenAI announced ChatGPT Atlas, an AI-native web browser with a built-in "Agent Mode" for task automation. 330 | - text: 1X announced Neo, a humanoid robot marketed as the first consumer-ready model for home use. 331 | special: true 332 | - date: November 2025 333 | info: 334 | - text: Moonshot AI released Kimi K2 Thinking, an open model setting new records in reasoning benchmarks. 335 | - text: OpenAI launched GPT 5.1, featuring specialized "Thinking" and "Instant" modes with expanded context. 336 | - text: xAI released Grok 4.1, combining high EQ with strong logic to top the LM Arena leaderboard. 337 | - text: Google debuted Gemini 3.0, a flagship "thinking" model that claimed the top spot on major benchmarks. 338 | - text: OpenAI introduced GPT 5.1 Codex Max, an agentic model built specifically for long-term coding tasks. 339 | - text: Google released Nano Banana Pro, a superior image generation and editing model based on Gemini 3. 340 | special: true 341 | - text: Anthropic announced Claude Opus 4.5, delivering elite coding and agentic performance at a significantly reduced price. 342 | special: true 343 | - text: Black Forest Labs launched FLUX 2, a high-performance open-weight image generation model. 344 | - text: DeepSeek released DeepSeekMath-V2 as open source, achieving gold-medal performance in math olympiads. 345 | special: true 346 | - text: Microsoft open-sourced Fara-7B, a small model optimized for browser agents and computer control. 347 | - text: Poetiq shatters the ARC-AGI-2 benchmark with a score of over 60%, surpassing the human average. 348 | --------------------------------------------------------------------------------