Model registry and catalog for AI inference across 376+ models from 58+ providers

Hanzo Models

Hanzo Models is the centralized model registry and catalog for all AI models available through the Hanzo platform. It provides model discovery, capability metadata, pricing information, and version management.

Production: https://models.hanzo.ai API: https://api.hanzo.ai/v1/models Pricing: https://pricing.hanzo.ai/v1/models

Overview

The model registry indexes 376+ models from 58+ providers, including 32 first-party Zen models. Every model served through api.hanzo.ai is registered here with standardized metadata.

Model Categories

Category	Count	Description
Zen (First-Party)	32	Hanzo-trained models running on Hanzo Engine
OpenAI	40+	GPT-4o, GPT-5 nano/mini, o1, o3 series
Anthropic	15+	Claude Haiku, Sonnet, Opus series
Meta	20+	Llama 3.x, Llama 4 series
Google	15+	Gemini 2.x, 3.x series
Mistral	10+	Mistral, Mixtral series
Other Providers	200+	Cohere, AI21, DeepSeek, Qwen, etc.

Model ID Format

All models follow the {provider}-{model-name} naming convention:

zen4-pro              # Zen 4 Pro (first-party)
zen4-mini             # Zen 4 Mini (first-party)
openai-gpt-5-nano     # OpenAI GPT-5 Nano
anthropic-claude-sonnet-4-5  # Anthropic Claude Sonnet 4.5
meta-llama-4-maverick-17b    # Meta Llama 4 Maverick 17B
google-gemini-2.5-flash      # Google Gemini 2.5 Flash

API

List Models

# OpenAI-compatible model listing
curl https://api.hanzo.ai/v1/models \
  -H "Authorization: Bearer $HANZO_API_KEY"

Response:

{
  "object": "list",
  "data": [
    {
      "id": "zen4-pro",
      "object": "model",
      "created": 1740000000,
      "owned_by": "hanzo",
      "capabilities": {
        "chat": true,
        "completion": true,
        "embedding": false,
        "image": false,
        "audio": false
      },
      "context_window": 128000,
      "max_output_tokens": 32768,
      "pricing": {
        "input": 0.000003,
        "output": 0.000015,
        "unit": "per_token"
      }
    }
  ]
}

Get Model Details

curl https://models.hanzo.ai/v1/models/zen4-pro

Model Capabilities

Each model exposes its capabilities:

Capability	Description
`chat`	Chat completions (multi-turn conversation)
`completion`	Text completions (single prompt)
`embedding`	Text embeddings (vector representations)
`image`	Image generation or understanding
`audio`	Speech-to-text or text-to-speech
`vision`	Image/video understanding
`tools`	Function calling / tool use
`json_mode`	Structured JSON output

Zen Models

Zen models are Hanzo's first-party models trained and served on Hanzo Engine. They are optimized for the Hanzo platform with native tool use, MCP integration, and low-latency inference.

Model	Parameters	Context	Specialty
`zen4-pro`	405B	128K	General purpose, reasoning
`zen4-mini`	70B	128K	Fast, cost-effective
`zen4-nano`	8B	32K	Edge deployment, low latency
`zen4-code`	70B	128K	Code generation and analysis
`zen4-vision`	70B	128K	Multimodal (text + image)

Zen models are built on Qwen3+ architecture and fine-tuned on Hanzo's proprietary datasets.

Provider Routing

When you send a request to api.hanzo.ai, the Gateway routes to the appropriate provider based on the model ID prefix:

Client -> api.hanzo.ai -> Gateway -> Provider
                                      ├── zen-*     -> Hanzo Engine
                                      ├── openai-*  -> OpenAI / DO AI
                                      ├── anthropic-* -> Anthropic / DO AI
                                      ├── meta-*    -> Meta / DO AI
                                      ├── google-*  -> Google / DO AI
                                      └── mistral-* -> Mistral / DO AI

Pricing

Model pricing is available through the Pricing API:

curl https://pricing.hanzo.ai/v1/models

Prices are denominated in USD per token. Input and output tokens are priced separately. Prices sync from upstream providers every 6 hours.

Model Selection

By Use Case

Use Case	Recommended Model	Why
General chat	`zen4-pro`	Best quality/cost ratio
Fast responses	`zen4-mini`	Low latency, good quality
Code generation	`zen4-code`	Optimized for code tasks
Image understanding	`zen4-vision`	Multimodal support
Embeddings	`text-embedding-3-small`	High quality, low cost
Budget	`zen4-nano`	Lowest cost per token

By Provider Preference

If you need a specific provider (compliance, licensing, etc.), use the provider prefix:

// Use Anthropic specifically
const response = await client.chat.completions.create({
  model: 'anthropic-claude-sonnet-4-5-20250514',
  messages: [{ role: 'user', content: 'Hello' }]
})

API gateway that routes model requests to providers

GPU inference engine serving Zen models

Real-time pricing for all models

Cloud platform with model management

Hanzo Models

On this page