unbiased ai

OpenClaw, thinking with Pareto

OpenClaw already lives in your chat apps and does the work. It just needs a model. Add Unbiased as a custom provider in openclaw.json, make unbiased/pareto the default, and every turn goes to Pareto, which selects frontier and open models for each request and blends their outputs into one answer. Your channels, skills and memory stay put.

Choose Personal · $10/month See the setup →

Tested live with OpenClaw 2026.10.550M tokens a week

A demonstration of OpenClaw on Pareto, recreated from a real run. Not a screenshot.

Add Pareto as a custom provider

OpenClaw reads its models from ~/.openclaw/openclaw.json, a JSON5 file. Add one provider under models.providers, list pareto as its model, and point the default agent at unbiased/pareto. Unbiased speaks the OpenAI Chat Completions format, so the provider’s api is openai-completions.

  1. Get an Unbiased API key

    Choose a plan, then create a key in the Unbiased platform. New accounts are reviewed and approved by a person before the key works.

  2. Put the key in your environment

    Export it as UNBIASED_API_KEY. OpenClaw fills ${UNBIASED_API_KEY} in the config from the environment, so the key never has to sit in the file.

  3. Add the provider block

    Paste the block below into openclaw.json. Keep /v1 on the baseUrl, and list pareto in the provider’s own models array: Unbiased does not publish a model list, so OpenClaw will not discover it. Set contextWindow too, or OpenClaw assumes 200,000 tokens.

  4. Check the config

    Run openclaw doctor. The config is strict, and an unknown key stops the Gateway from starting, so doctor is the quick way to catch a typo.

  5. Send it a task

    Message OpenClaw from any channel you have connected, or run one headless turn with openclaw agent exec, the command from our live test.

Tested on October 20, 2026OpenClaw 2026.10.5 on Pareto (pareto). It found the off-by-one in a failing Python test, edited the file and reran the test to green. Run record.
~/.openclaw/openclaw.json
{
  "agents": {
    "defaults": { "model": { "primary": "unbiased/pareto" } }
  },
  "models": {
    "mode": "merge",
    "providers": {
      "unbiased": {
        "baseUrl": "https://api.unbiased.ai/v1",
        "apiKey": "${UNBIASED_API_KEY}",
        "api": "openai-completions",
        "models": [
          {
            "id": "pareto",
            "name": "Pareto",
            "contextWindow": 1000000,
            "maxTokens": 131072
          }
        ]
      }
    }
  }
}
Shell
export UNBIASED_API_KEY=your-unbiased-api-key
openclaw doctor
# One headless turn, as in our live test
openclaw agent exec --model unbiased/pareto \
  --cwd ./proj \
  "test_chunk.py fails. \
Fix the bug in chunk.py, \
then run python3 test_chunk.py to confirm."

"mode": "merge" (the default) keeps your other providers alongside Unbiased. To keep your config somewhere else, point OPENCLAW_CONFIG_PATH at that file.

The lobster stays. The model changes.

One provider block and one default. Everything OpenClaw does around the model carries on as before.

One model string, several models behind it

Pareto is Unbiased’s blended AI model. On each request it selects frontier and open models and blends their outputs into one answer, so OpenClaw talks to one model string, pareto, and never has to pick.

“When the frontier improves, Pareto improves; when frontier prices fall, you should benefit.” From the pricing page.

How a Pareto answer gets made →
An illustration, not a wiring diagram. The blend includes frontier and open models, and its composition can change.

What an agent task costs on Pareto

OpenClaw runs real tasks with real tools, so the price of a finished task matters more than the price of a token. The model card publishes both for Pareto on two public agent benchmarks.

Terminal-Bench 4.0Tasks an agent completes in a terminal
$0.48/task
50.8% solved
DeepSWE v1.1Changes an agent makes in a real codebase
$0.24/task
69.9% solved

Pareto 26.10, preliminary October 1, 2026 runs that may change; denominators and the cost method still require confirmation. Cost is the reported mean USD per task at API rates, not an estimate for your workload. Model card and method.

Pay for OpenClaw by the month

One API key works across OpenClaw and any other tool you point at Pareto. Pick a monthly allowance, or pay per token on Team.

Personal

$10/month

  • 50M tokens a week
  • 20 requests per minute
  • For one person
Choose Personal

Personal Max

$100/month

  • 600M tokens a week
  • 60 requests per minute
  • For one person
Choose Personal Max

Team

Pay as you go

  • Prepaid credits at published rates
  • 300 requests per minute
  • No weekly token cap
Get started with Team

Subscriptions are for one person and are not unlimited. New API accounts are reviewed by a person before access turns on, and commercial or production use needs a separate written agreement. Rates and limits are on Pricing.

OpenClaw on Pareto, questions answered

Does OpenClaw work with Pareto?

Yes. We ran OpenClaw 2026.10.5 against Pareto on October 20, 2026 with the config on this page: it read a failing Python test, edited the file and reran the test to green, using its read, edit and exec tools. Format compatibility is not a promise that every OpenClaw feature works, so tell us if one doesn’t.

Why is the model called unbiased/pareto and not just pareto?

OpenClaw names models as provider/model. unbiased is the provider key you chose in models.providers, and pareto is the model ID. If you name the provider something else, change the ref to match.

I added pareto, but OpenClaw says the model is unknown. Why?

It has to be listed in models.providers.unbiased.models. An entry under agents.defaults.models only sets aliases and metadata; it does not register a model. Run openclaw doctor after the change.

Do I need an OpenAI or Anthropic key as well?

No. With this config OpenClaw sends model requests to Unbiased with your Unbiased key, and your Unbiased plan pays for them. OpenClaw itself is free and open source.

Can I switch back to my old model?

Yes. Set agents.defaults.model.primary back to your previous model ref, or pass --model for a single turn. Leave the Unbiased block in place if you want Pareto one edit away.

Is Pareto made by the OpenClaw team?

No. Pareto is Unbiased’s model. OpenClaw is made by the OpenClaw Foundation, and Unbiased is not affiliated with or endorsed by them. You keep OpenClaw and point it at the Unbiased API.

What happens when I hit the weekly allowance?

Requests over your plan’s limits return an HTTP 429 error. Personal includes 50M tokens a week and 20 requests per minute. Personal Max raises that to 600M a week and 60 requests per minute for $100/month. Team uses prepaid pay-as-you-go credits with no weekly token cap.

Do I choose between models?

No. You send every request to one model string, pareto. Pareto selects frontier and open models for each request and blends their outputs into one answer.

Is my code kept?

Personal and Personal Max are not our zero data retention tier: prompts and responses may be kept up to 30 days, as section 5(c) of the Terms explains. For work that must not be kept, ask about the Zero Data Retention API on the Private AI page.

Pareto works in other tools too

The same key and the same model string go into each of these. See them all.

Product names belong to their owners. Except where a guide says otherwise, Unbiased is not affiliated with or endorsed by the makers of the tools on this page.

Give OpenClaw a model and a chore

One provider block, one key and the assistant you already message. Pareto starts at $10 a month for 50M tokens a week.

Choose Personal · $10/month See the setup →