Zen Models

Zen LM by Hanzo AI is a comprehensive model family built using Zen MoDE (Mixture of Diverse Experts) — diverse expert specializations fused into a unified, high-performance family.

All API models are accessed at api.hanzo.ai. API key prefix: hk-. Open-weight models are available on HuggingFace.

Zen 5 — Next Generation

Next-generation agentic models with native chain-of-thought, currently in preview. Training on-chain via NVIDIA TEE on hanzo.network.

Request early access: z@hanzo.ai

Zen 4 — Production API

Latest generation production models with MoDE architecture. The recommended choice for all new applications.

Code

Specialized models for code generation, review, debugging, and agentic programming.

Zen 3 — Previous Generation API

Earlier generation API models — language, vision, multimodal, and safety. Still production-ready and widely deployed.

Embedding & Retrieval

Text embeddings and search reranking via /v1/embeddings and /v1/rerank.

Image Generation

Text-to-image generation via /v1/images/generations.

Audio & Speech

Speech-to-text, text-to-speech, and streaming ASR via /v1/audio/.

Foundation — Open Weights

General-purpose open-weight models available on HuggingFace in GGUF, SafeTensors, MLX, and ONNX formats.

Vision — Open Weights

Vision-language and multimodal open-weight models.

Safety & Guardrails

Content moderation and safety guardrail models.

Agents

Agent-optimized models for tool use, planning, and autonomous workflows.

Video (Coming Soon)

Audio Creative (Coming Soon)

3D & Spatial (Coming Soon)

API Usage

from hanzoai import Hanzo

client = Hanzo(api_key="hk-your-api-key")

response = client.chat.completions.create(
    model="zen4",
    messages=[{"role": "user", "content": "Hello!"}],
)
print(response.choices[0].message.content)

import Hanzo from '@hanzo/ai';

const client = new Hanzo({ apiKey: 'hk-your-api-key' });

const response = await client.chat.completions.create({
  model: 'zen4',
  messages: [{ role: 'user', content: 'Hello!' }],
});
console.log(response.choices[0].message.content);

See API Reference for full documentation.

Open Weights Formats

All locally-runnable models are available in multiple formats: