- Full redesign with header, footer, pricing, chat widget, Try Zen CTA
- All 49 models listed across 10 modalities
- 30 new model documentation pages (44 total)
- Hanzo SDK examples (hanzoai) replacing OpenAI SDK
- Ecosystem, research, and open weights sections
- Login (hanzo.id) and Try Zen (hanzo.chat) in header
Rebuilt from hanzoai/docs monorepo. All upstream model references
(Qwen3, GLM, Zhipu) removed from public docs. zen-max now correctly
shows 1.04T MoE specs.
Replace stale standalone site with zen-docs built from the
hanzoai/docs monorepo (apps/zen-docs). 32 static pages including
model docs, API reference, training guides, and datasets.
All model data sourced from @zenlm/models npm package.
- Homepage: Zen4 Consumer, Coder, and Ultra model tables
- Models page: Full Zen4 lineup with cards and comparison table
- Added datasets page
- Status labels: Available, Cloud Only, Coming Soon
- All HuggingFace and GitHub links for zenlm/ repos
- Based on Moonshot AI Kimi K2 Thinking (DeepseekV3)
- 384 experts, 8 active per token (~14B active parameters)
- 256K context with 96K-128K thinking tokens
- Top benchmarks: AIME 99.1%, SWE-Bench 71.3%, BrowseComp 60.2%
- HuggingFace: https://huggingface.co/zenlm/zen-max
- Updated zen-omni to show Qwen3-Omni as base (multimodal: text+vision+audio)
- Added docsLink and paperLink to ModelCard component
- zen-omni now shows proper modalities and type
- Added links to documentation sites and whitepapers
- Updated all model cards with complete resource links
- Fixed .gitignore to only ignore root-level directories
- Changed models/ to /models/ (and similar for other paths)
- This allows docs/models/ and src/app/models/ to be tracked
- Added missing files:
- src/app/models/page.tsx (source code)
- docs/models/index.html (static page)
- docs/_next/static/chunks/app/models/ (Next.js chunks)
- Models page is now properly included in repository
- Convert to Next.js 15 with TypeScript
- Create reusable React components for Header/Footer (DRY)
- Convert all pages to TSX: page.tsx, models/page.tsx, research/page.tsx
- Configure static export to docs/ for GitHub Pages
- Add .nojekyll for proper routing
- All builds generate static HTML (SSG)
- Remove old build.js template system
- TypeScript for type safety
- Next.js Image optimization disabled for static export
- Create shared header/footer templates in docs/templates/
- Add build.js script to compile templates into final HTML
- Add package.json with build script
- Create src/ directory for source templates
- Update index.html to use template system
- No React/Next.js - keeping it simple for 3-page static site
- Link Zen LM logo to home page across all pages
- Move GitHub/HuggingFace icons to far right of header
- Use HuggingFace PNG logo instead of SVG
- Update parameter range from 0.6B-32B to 0.6B-1T+
- Improve header layout with icons on the right
- Ensure consistent navigation across all pages
MAJOR UPDATE: Full website transformation for XR/VR/Robotics focus
Pages Added:
- models.html: Complete catalog of 24+ models organized by category
- research.html: 15 compiled research papers with download links
Papers Built & Added (15 PDFs in docs/papers/):
- zen-technical-report.pdf (master technical report)
- Core: zen-eco, zen-pro, zen-coder, zen-agent, zen-guard
- Multimodal: zen-omni, zen-designer, zen-artist, zen-musician, zen-director, zen-foley, zen-scribe, zen-world
- Infrastructure: gym
Landing Page Enhanced (index.html):
- Mission: Real-Time Hyper-Modal AI for XR/VR/Robotics
- Emphasizes sub-10ms latency, spatial awareness, multimodal fusion
- Stats dashboard: 24+ models across language/vision/audio/3D/video
- Streamlined overview (detailed models → models.html)
- CTAs to Models and Research pages
README Updated:
- Complete documentation of site structure
- All 24 models listed with descriptions
- Development & deployment instructions
- Citation information
- XR/VR/Robotics mission front and center
Navigation:
- Consistent across all pages: Home | Models | Research
- Active states for current page
Technical:
- All LaTeX papers compiled with pdflatex
- PDFs organized in docs/papers/
- GitHub Pages auto-deployment configured
- Clean separation: homepage (overview) → models page (catalog) → research page (papers)
- Remove all model files/dirs from repo (moved to parent)
- Add GitHub and HuggingFace icons to header
- Remove card animation for immediate visibility
- Update copyright to Zen Authors
- Enhanced .gitignore to prevent future clutter
- Repo now contains only docs site and workflows
- Add GitHub Actions workflow for automated Pages deployment
- Change all branding from Zen AI to Zen LM
- Remove Japanese character (禅) from site
- Update meta tags and page titles
- Ready for deployment to zenlm.org
- Comprehensive docs for all 24 zen models
- Core models: nano, eco, omni, pro, coder, next
- Multimodal: designer, artist, video, 3d, world, voyager, musician, foley, scribe, director
- Specialized: agent, code, guard, embedding
- Infrastructure: engine, gym, family, blog
- Black theme optimized for researchers
- Official zen logo and favicon
- Clean repo with no large files or model weights
Hanzo Dev