AI research and infrastructure
as open, public goods.
Hanzo builds open-weight models, cloud infrastructure, and agent frameworks — freely available to researchers, developers, and the broader AI ecosystem. 500+ open source repos. MIT and Apache licensed. 25% of compute revenue goes back to OSS contributors.
Building the Future of Safe AI
From frontier model development to edge deployment, we're advancing AI capabilities while prioritizing safety, privacy, and human alignment at every step
Frontier AI Models
State-of-the-art language models and multimodal AI — the Zen family spans 600M to 1T+ parameters across text, vision, audio, video, and 3D
AI Safety & Alignment
Research and implementation of AI safety measures, interpretability tools, and alignment techniques across all Hanzo products
AI Workforce & Automation
Deploy an autonomous AI team — Hanzo Bot gives you 16 specialized AI agents for engineering, design, marketing, and operations
Platform & Infrastructure
Enterprise-grade AI platform with LLM gateway supporting 200+ models, secure deployment, and resilient infrastructure — from edge to cloud
Research at the Frontier
58 published papers across AI, cryptography, consensus, and distributed systems. Our research teams investigate the safety, efficiency, and societal impacts of AI.
AI & Machine Learning
100+ model weights from 0.6B to 1T+ with Training-Free GRPO optimization
Cryptography & Security
Post-quantum cryptography and GPU-accelerated fully homomorphic encryption
Consensus & Networks
Multi-consensus blockchain architecture with TEE attestation
Zen AI Models
Zen4/5 — from 4B to 2T+ MoDE parameters with GT-QLoRA MoE fine-tuning
Latest Publications
Training-Free Adaptation via Active Semantic Optimization
Achieving 18.2% SWE-bench resolution with zero additional training through TF-GRPO and BitDelta compression.
Quasar: Dual-Certificate Quantum-Secure Consensus
Post-quantum finality using BLS and Corona threshold signatures for blockchain security.
Hamiltonian Market Maker for Decentralized AI Compute
Novel invariant H(Ψ,Θ) = κ achieving <200ms latency and 98.7% price stability for AI compute markets.
Investors & Partners
Backed by world-class partners
Trusted By
易經 · The Engineering Philosophy
Ten principles drawn from the 64 hexagrams of the I-Ching. Orthogonality. Smallness. Completeness. Clarity. Composability.
Ancient pattern language for systems that last. The same framework behind every model, protocol, and line of code we ship.
The Zen of HanzoLet's Build Together
From research collaborations to enterprise AI deployments, we partner with teams pushing the boundaries of what's possible
Research Division
Partnerships, commercial terms and enterprise enquiries.
hello@hanzo.industriesCommercial Division
Partnerships, commercial terms and enterprise enquiries.
hello@hanzo.industriesCorporate Headquarters
1824 S. Fairfax Ave
Los Angeles, CA 90019
United States
Secure Line
+1 (913) 777-4443
Available 24/7