Skip to content
Language: English (United States)

Every model, and what it costs

Priced

Dollars per million tokens, in and out, from the live catalog.

35 models.

    • Enso Auto

      enso-auto

      Adaptive Enso router — selects the right Enso model per request by reasoning depth and task, billing at the model that answers.

      text

      1M

      $0.330

      $1.65

    • Enso Flash

      enso-flash

      Fast, economical Enso model for high-volume, low-latency everyday chat, coding, and tool calls.

      text

      262K

      $0.330

      $1.65

    • Enso Pro

      enso-pro

      Balanced frontier reasoning and coding model for production agents, tool use, and repo-scale engineering.

      text

      1M

      $3.30

      $16.50

    • Enso Max

      enso-max

      Extended frontier reasoning model for deep multi-step architecture, research, and autonomous coding loops.

      text

      1M

      $4.30

      $21.50

    • Enso Ultra

      enso-ultra

      Adaptive fan-out — probes a task-appropriate model, escalates to a top-K panel only when needed, then verifies and selects the best answer.

      text

      1M

      $5.24

      $26.18

  • Zen models

    14 models

    Open weights, made with Zoo Labs Foundation. Download every Zen model at huggingface.co/zenlm, or call it here.

    Run Zen: the docs →
  • Kai decision models

    3 modelsKai: decisions and their benchmarks →
    • Kai

      kai

      Hanzo Kai is a structured decision model answering typed questions at /v1/decisions with calibrated probabilities and zero output tokens. Full multimodal decision architecture (~850M parameters, 32K reach); native vision and audio towers: Kai-2 target.

      TextStateVision (Kai-2 target)

      33K

      $0.021

      $0.00

    • Decision Benchmarks

      11 suites won vs jev

      Ahead of Jev on 11 suites: AG News, DAIR Emotion, Banking77, Support triage, Email spam, Phishing, Toxicity, RAG relevance, Model routing, Typed decisions, MASSIVE, 51 languages. 84.4% mean accuracy across these 11 suites, against Jev's 76.6%. 100.0% of repeated requests answer identically, against Jev's 44.2%.

      benchmarks

      —

      —

      —

    • Autonomous Agent Steering

      decision-ready

      A choice, a score or a yes/no, each with its probabilities, for agent loops such as Hanzo Dev's.

      choicescorenoul

      —

      —

      —

  • Creative models

    6 models

    Video, images, 3D, sound effects, voice and music, each its own model, in research preview.

    Make something in Design →
    • Koma

      koma-1api.hanzo.ai/v1/video

      Cinematic video generation, character continuity, motion, animation, and shot-level creative control.

      Research previewVideo

      —

      —

      —

    • Iro

      iro-1api.hanzo.ai/v1/images

      High-fidelity image generation, creative direction, editing, character design and consistent visual assets.

      Research previewImage

      —

      —

      —

    • Mori

      mori-1api.hanzo.ai/v1/3d

      3D assets, geometry, materials, environments and spatially consistent scenes.

      Research preview3D

      —

      —

      —

    • Oto

      oto-1api.hanzo.ai/v1/foley

      Film Foley, immersive sound effects, environmental audio and synchronized scene sound.

      Research previewFoley

      —

      —

      —

    • Koe

      koe-1api.hanzo.ai/v1/audio/speech

      Directable speech synthesis, character performances, dubbing and multilingual dialogue.

      Research previewVoice

      —

      —

      —

    • Uta

      uta-1api.hanzo.ai/v1/music

      Soundtracks, original songs, musical ideas, cue-driven scoring and creative music production.

      Research previewMusic

      —

      —

      —

  • Flagships from other labs

    7 modelsEvery model, priced →
    • Claude Fable 5

      anthropic/claude-fable-5

      Claude Fable 5 is a Mythos-class model from Anthropic, built for autonomous knowledge work and coding. It supports text, image, and file inputs with text output, with reasoning support and...

      text

      1M

      —

      —

    • GPT-3.5 Turbo (older v0613)

      openai/gpt-3.5-turbo-0613

      GPT-3.5 Turbo is OpenAI's fastest model. It can understand and generate natural language or code, and is optimized for chat and traditional completion tasks. Training data up to Sep 2021.

      text

      4K

      —

      —

    • Gemini 2.5 Pro

      google/gemini-2.5-pro

      Gemini 2.5 Pro is Google’s state-of-the-art AI model designed for advanced reasoning, coding, mathematics, and scientific tasks. It employs “thinking” capabilities, enabling it to reason through responses with enhanced accuracy...

      text

      1M

      $1.39

      $11.13

    • DeepSeek V3

      deepseek/deepseek-chat

      DeepSeek-V3 is the latest model from the DeepSeek team, building upon the instruction following and coding abilities of the previous versions. Pre-trained on nearly 15 trillion tokens, the reported evaluations...

      text

      164K

      $0.356

      $0.991

    • Grok 4.20

      x-ai/grok-4.20

      Grok 4.20 is a reasoning model from SpaceXAI with industry-leading speed and agentic tool calling capabilities. It combines the lowest hallucination rate on the market with strict prompt adherance, delivering...

      text

      2M

      —

      —

    • Muse Glimmer 30B

      meta/muse-glimmer-30b

      Muse Glimmer 30B is a dense, open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark and optimized for autonomous agents on consumer hardware. It is suited for long-horizon...

      text

      131K

      $0.334

      $1.34

    • Command A

      cohere/command-a

      Command A is an open-weights 111B parameter model with a 256k context window focused on delivering great performance across agentic, multilingual, and coding use cases. Compared to other leading proprietary...

      text

      256K

      $2.78

      $11.13

Hanzo19 models.

How models are called and billed →
Every model
Enso
Aion Labs
Alibaba
Amazon
Anthracite
Anthropic
assemblyai
black-forest-labs
ByteDance
canopylabs
Cognitive Computations
Cohere
deepgram
DeepSeek
dots-studio
elevenlabs
elorian
Hanzo
fireworks
fish-audio
Google
Gryphe
hexgrad
heygen
IBM Granite
Inception
inclusionAI
inference-net
kwaivgi
Liquid AI
Mancer
Meituan
Meta
Microsoft
Minimax
Mistral
moondream
Moonshot AI
Morph
Nex-AGI
Nous Research
NVIDIA
OpenAI
OpenRouter
particle-ai
Perceptron
Perplexity
prism-ml
Qwen
recursive
reflection
Reka AI
Relace
ricursiveai
runway
Sakana AI
Sao10K
scaledown
sesame
stealth
StepFun
tbc
Tencent
TheDrummer
Thinking Machines
typesafe
unbiased
Undi95
Upstage
Writer
xAI
Xiaomi
xingchen-agi
Z.ai

Enso routes. Kai decides. Zen reasons.

Three families

  1. routes

    Enso

    Picks the model for every call, in every modality.

    Explore Enso →
  2. decides

    Kai

    Answers a choice, a score or a yes/no, with the probability of each.

    11 · Suites won · Kai vs Jev

    Explore Kai →
  3. reasons

    Zen

    Open weights for the code, the copy, the images and the voice.

    Zen 6 and Zen 6 Flash now · Zen 7 in research preview

    Explore Zen →

Proprietary orchestration, on open foundations

Enso vs Zen

Enso

Proprietary · Hanzo Cloud only

Zen

Open weights · Zoo Labs Foundation

Flash, Pro and Ultra presets
Free to download and self-host
OpenAI- and Anthropic-compatible
Or call it on Hanzo Cloud
Metered and audited per organization
From a 501(c)(3) non-profit

Ahead of Jev on 11 suites

Decision Benchmarks

  • AG News
  • DAIR Emotion
  • Banking77
  • Support triage
  • Email spam
  • Phishing
  • Toxicity
  • RAG relevance
  • Model routing
  • Typed decisions
  • MASSIVE, 51 languages
Kai: decisions and their benchmarks →

11

Suites won · Kai vs Jev

84.4%

mean accuracy across these 11 suites

against Jev's 76.6%

100.0%

of repeated requests answer identically

against Jev's 44.2%

Try any model

Test any model directly on the gateway. Switch models by changing one string.

Sign in to chat with Enso, Zen, or any model on the endpoint. Pick a model, start a thread, billed to your account and kept private.

Live from api.hanzo.aiPOST /v1/chat/completionsexample
{ "model": "enso-auto", "messages": [{ "role": "user", "content": "Who are you?" }] }

Hanzo's adaptive frontier router — selects the right Enso model per request with 1M-context overflow.

1Mtext

Your first request

The chat-completions shape you already send, at api.hanzo.ai/v1, in the language you write.

cURL
curl https://api.hanzo.ai/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $HANZO_API_KEY" \
  -d '{
    "model": "enso-auto",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

Frontier AI families

Proprietary and open weight models for text, decisions, video, image, and more — one account, one API key.