Models
All models available in Hanzo Chat — 14 Zen models and 100+ third-party from OpenAI, Anthropic, Google, Meta, Mistral, Together, and Groq.
Models
Hanzo Chat provides access to 14 first-party Zen models and 100+ third-party models through the Hanzo LLM Gateway.
Zen Models (First-Party)
All 14 Zen models are available by default. These are Hanzo's own frontier models built on MoDE (Mixture of Distilled Experts) architecture.
| Model | Type | Context | Input $/MTok | Output $/MTok |
|---|---|---|---|---|
| zen4 | Flagship | 128K | $3.00 | $9.60 |
| zen4-ultra | Max Reasoning | 128K | $3.00 | $9.60 |
| zen4-pro | High Capability | 128K | $2.70 | $2.70 |
| zen4-max | Large Documents | 1M | $3.60 | $3.60 |
| zen4-mini | Fast & Efficient | 128K | $0.60 | $0.60 |
| zen4-thinking | Chain-of-Thought | 128K | $2.70 | $2.70 |
| zen4-coder | Code Generation | 128K | $3.60 | $3.60 |
| zen4-coder-pro | Premium Code | 128K | $4.50 | $4.50 |
| zen4-coder-flash | Fast Code | 128K | $1.50 | $1.50 |
| zen3-omni | Multimodal | 128K | $1.80 | $6.60 |
| zen3-vl | Vision-Language | 32K | $0.45 | $1.80 |
| zen3-nano | Edge | 32K | $0.30 | $0.30 |
| zen3-guard | Content Safety | 8K | $0.30 | $0.30 |
| zen3-embedding | Embeddings | 8K | $0.39 | -- |
Model Recommendations
| Use Case | Recommended Model |
|---|---|
| General chat | zen4 or zen4-pro |
| Code generation | zen4-coder or zen4-coder-pro |
| Quick responses | zen4-mini |
| Image analysis | zen3-omni or zen3-vl |
| Deep reasoning | zen4-thinking or zen4-ultra |
| Long documents | zen4-max (1M context) |
| Budget-friendly | zen3-nano ($0.30/MTok) |
| Content moderation | zen3-guard |
Third-Party Models
100+ models from all major providers are available through the Hanzo Gateway:
OpenAI
| Model | Type |
|---|---|
| gpt-4o | Flagship multimodal |
| gpt-4o-mini | Fast, affordable |
| o3 | Advanced reasoning |
| o4-mini | Fast reasoning |
Anthropic
| Model | Type |
|---|---|
| claude-opus-4 | Most capable |
| claude-sonnet-4.5 | Best balance |
| claude-haiku-4.5 | Fast, efficient |
| Model | Type |
|---|---|
| gemini-2.5-pro | Advanced reasoning |
| gemini-2.5-flash | Fast multimodal |
Meta
| Model | Type |
|---|---|
| llama-4-maverick | 400B MoE |
| llama-4-scout | 109B MoE |
| llama-3.3-70b | Open weights |
Mistral
| Model | Type |
|---|---|
| mistral-large | Flagship |
| codestral | Code-specialized |
| mistral-small | Efficient |
Together AI
50+ open models including Llama, Mixtral, Qwen, DeepSeek, and more with fastest inference.
Groq
Ultra-fast inference for Llama, Mixtral, and Gemma models.
Configuring Models
Via chat.yaml
endpoints:
hanzo:
baseURL: "https://api.hanzo.ai/v1"
apiKey: "${HANZO_API_KEY}"
models:
default:
- zen4
- zen4-coder
- zen4-mini
fetch: true # Auto-discover additional models
openAI:
apiKey: "${OPENAI_API_KEY}"
models:
default:
- gpt-4o
- o3Auto-Discovery
Set fetch: true on any endpoint to automatically discover available models from the API. This is useful when new models are added — they appear without config changes.
Custom Display Names
endpoints:
hanzo:
modelDisplayLabel: "Zen"
iconURL: "https://cdn.hanzo.ai/img/logo-white.svg"Pricing
- $5 free credit on every new account (expires in 30 days)
- Prepaid billing — add credits at console.hanzo.ai
- $1 minimum balance required for API calls
- No surprise bills — service stops when credits are depleted
See the full pricing table at zenlm.org/pricing.
How is this guide?
Last updated on