mirror of
https://github.com/hanzoai/papers.git
synced 2026-10-22 17:20:17 +00:00
zeekay
2773271544
35+ papers documenting Hanzo AI platform evolution covering: - Crowdstart cloud-native commerce (2014 era) - Astle.js reactive SDK, Verus analytics (2015 era) - Earle genetic marketing optimization (2016 era) - Hanzo AI-first commerce platform (Techstars 2017) - Checkout, analytics ML (2018 era) - API gateway, SDK ecosystem, webhooks (2019 era) - MCP protocol, Platform PaaS (2021 era) - Agent SDK, ML framework, search, identity NFT (2022 era) - Jin multimodal AI, DSO protocol (2023 era) - Candle Rust ML, ACI decentralized compute (2024 era) - Chat, Flow, Operative, Network whitepaper, ASO, HMM, LLM gateway, Zen model family (2025 era) Papers describe historical work from 2014-2025. All compiled into this repository October 2025 - March 2026. Note: Previous git history had fabricated per-commit timestamps claiming individual commits from 2014-2025. This corrects the record — all content was committed in a single session.
1.4 KiB
1.4 KiB
Zen AI Model Family
Repository: https://github.com/zenlm/papers
Organization: ZenLM
Co-developed by: Hanzo AI Inc (Techstars '17) & Zoo Labs Foundation (501c3)
Zen AI is a comprehensive family of ultra-efficient language models that serve as the base underlying frontier models for both Hanzo AI and Zoo Labs projects. All models are based on Qwen3 architectures with innovative optimizations.
Main Documentation
- Main Whitepaper: ZEN_WHITEPAPER_2025.md (v1.0.1)
- Technical overview of complete Zen ecosystem
- 600M to 480B parameter models
- 95% reduction in energy consumption
- Recursive AI Self-Improvement System (RAIS) with 94% effectiveness
Model Papers
Visit https://github.com/zenlm/papers for complete documentation including:
- Core Models: Nano (600M edge), Eco (balanced), Coder (code gen), Omni (multimodal), Next (MoE), Guard (safety)
- Specialized Models: Artist, Designer-Instruct, Designer-Thinking, Scribe, Director, and more
- Architecture Overview: zen_family_overview.tex
All Zen models leverage Qwen3 base architectures with:
- Grouped Query Attention (GQA) - 75-87.5% KV cache reduction
- SwiGLU Activation - 10-15% improvement over ReLU
- RMSNorm - 20% faster than LayerNorm
- RoPE - Contexts up to 128K tokens
- Mixture of Experts (MoE) - Up to 96.25% parameter efficiency