Hanzo AI
Try Hanzo
Enso · Hanzo’s agentic model

The right model for every call.

A learned router behind one OpenAI- and Anthropic-compatible endpoint.

Try Hanzo See benchmarks

Drop-in

Keep your client.

Same key on /v1/chat/completions and /v1/messages. Enso picks among 600+ models.

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

Flash · Pro · Ultra

Pick speed or depth per call.

GPQA-Diamond, measured by Hanzo. Rates per million tokens, in → out.

Enso Ultra

enso-ultra
98.0%GPQA-Diamond · 194 of 198
$5 → $25For work where a wrong answer is expensive: research, security review and long autonomous runs.

Enso Pro

Default
enso
96.0%GPQA-Diamond · 190 of 198
$4 → $20For coding, code review and agents that answer while someone waits.

Enso Flash

enso-flash
92.9%GPQA-Diamond · 184 of 198
$2 → $4For chat, extraction, classification and simple agent steps at volume.

Accuracy at cost

The top score, for less.

Enso Ultra scores 98.0% at $25 per million output tokens. gpt-5.2-pro charges $139 and scores 93.2%.

76828894100$0.5$8$120GPQA %cheap ← output $/MTok → expensivegpt-5.5 — 93.6% GPQA-Diamond · $8.25/MTok (vendor-reported)gpt-5.5 93.6%gpt-5.2-pro — 93.2% GPQA-Diamond · $138.6/MTok (vendor-reported)gpt-5.2-pro 93.2%gpt-5.6-sol — 92.9% GPQA-Diamond · $25/MTokgpt-5.6-sol 92.9%kimi-k2.6 — 89.1% GPQA-Diamond · $2.71/MTok (vendor-reported)kimi-k2.6 89.1%qwen3.5-397b-a17b — 88.4% GPQA-Diamond · $2.04/MTok (vendor-reported)qwen3.5-397b-a17b 88.4%opus-4.8 — 87.4% GPQA-Diamond · $21/MTokopus-4.8 87.4%glm-5.2 — 85.6% GPQA-Diamond · $3.73/MTok (vendor-reported)glm-5.2 85.6%gemma-4-31b — 84.3% GPQA-Diamond · $0.44/MTok (vendor-reported)gemma-4-31b 84.3%fable-5 — 81.3% GPQA-Diamond · $42/MTokfable-5 81.3%enso-ultra — 98.0% GPQA-Diamond · $25/MTokenso-ultra 98.0%enso — 96.0% GPQA-Diamond · $20/MTokenso 96.0%enso-flash — 92.9% GPQA-Diamond · $4/MTokenso-flash 92.9%
768084889296100$0.5$2$8$30$120GPQA %cheap ← output $/MTok → expensivegpt-5.5 — 93.6% GPQA-Diamond · $8.25/MTok (vendor-reported)gpt-5.5 93.6%gpt-5.2-pro — 93.2% GPQA-Diamond · $138.6/MTok (vendor-reported)gpt-5.2-pro 93.2%gpt-5.6-sol — 92.9% GPQA-Diamond · $25/MTokgpt-5.6-sol 92.9%kimi-k2.6 — 89.1% GPQA-Diamond · $2.71/MTok (vendor-reported)kimi-k2.6 89.1%qwen3.5-397b-a17b — 88.4% GPQA-Diamond · $2.04/MTok (vendor-reported)qwen3.5-397b-a17b 88.4%opus-4.8 — 87.4% GPQA-Diamond · $21/MTokopus-4.8 87.4%glm-5.2 — 85.6% GPQA-Diamond · $3.73/MTok (vendor-reported)glm-5.2 85.6%gemma-4-31b — 84.3% GPQA-Diamond · $0.44/MTok (vendor-reported)gemma-4-31b 84.3%fable-5 — 81.3% GPQA-Diamond · $42/MTokfable-5 81.3%enso-ultra — 98.0% GPQA-Diamond · $25/MTokenso-ultra 98.0%enso — 96.0% GPQA-Diamond · $20/MTokenso 96.0%enso-flash — 92.9% GPQA-Diamond · $4/MTokenso-flash 92.9%
Solid ring: measured by Hanzo. Dashed: reported by the vendor.Every benchmark →

Hanzo Cloud

Managed, metered, audited.

Billed per organization

Pay the preset’s rate per token. Nothing stacked.

Traced

Every response records which models ran, in the console and the API.

Your pool

Opt providers or models out to meet data and compliance rules.

Private by default

Your inputs and outputs train nothing unless you opt in.
Pricing →
Enso is one of three Hanzo model families. Compare them →

Build on Enso.

Try HanzoRead the docs