Pareto / Agent and automation builders
More workflow. Less API spend.
Run Pareto in your agent workflow. One model string, published API rates, and prepaid credits to keep your evaluation bounded.
Team · prepaid API credits
$0.80/ million input tokens
$0.03 cached input · $3.20 output / million tokens
300 requests/minute · no weekly token cap
New API accounts need manual approval. Payment does not mean instant access.
One model string.
Models compete for your request.
Pareto is Unbiased’s blended AI model: several frontier and open models answer every request and the best result is returned.
It is built so AI providers compete for every request. See preliminary benchmarks and cost per task →
Built around your work
Test a workflow you already run.
Cost per completed run
Divide spend by finished workflows, not by API calls.
Tools and formats under test
Point your automation at the documented OpenAI-compatible endpoint and run your real tool-calling steps.
Limits you can plan around
Team pay-as-you-go has no weekly token cap.
Get connected
Start small.
Try your real work.
Choose your plan, get approved, then connect. Onboarding help is optional and subject to availability.
Request a walkthroughThe documented API route uses https://api.unbiased.ai/v1 with model pareto. OpenAI Chat Completions and Anthropic Messages formats are documented. Validate your client's streaming, tools and error handling; format compatibility is not a promise that every integration works.
Documented API connection
curl https://api.unbiased.ai/v1/chat/completions \
-H "Authorization: Bearer $UNBIASED_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"pareto","messages":[{"role":"user","content":"Hello"}]}'Confirm permitted use before commercial or production deployment. Terms · Usage Policy · Data Policy
Illustrative token math, not a measured customer result. Assumes 1M uncached input tokens, 1M cached input tokens and 0.25M output tokens across a batch. Categories are separate; no token is counted twice. Actual use, retries, cache eligibility and other costs change the bill. Divide by your own completed runs or accepted results for a per-result figure.
What can I buy today?
Personal is $10/month with 50M tokens a week and 20 requests per minute. Personal Max is $100/month with 600M tokens a week and 60 requests per minute. Team uses prepaid pay-as-you-go credits, with no weekly token cap and 300 requests per minute. Subscriptions are for one person and are not unlimited.
Will tool calling work with my automation tool?
The documented API exposes OpenAI Chat Completions and Anthropic Messages formats, so tools that speak those protocols can connect. Every automation platform has not been tested; validate triggers, tool schemas and error handling on your own instance before relying on them.
Will failures and retries increase?
Measure it on your workflow. Run both providers for a defined period and compare completed runs, retry counts and latency. Retries bill as input and output tokens, so they belong in the cost math.
Is this an inference API or an automation builder?
Pareto is an inference API plus a macOS coding workspace. It does not replace your workflow orchestrator; validate the connection before using it in unattended workflows.
When can I start using the API?
Create or sign in to your platform account to choose a plan. New API accounts require manual approval before access turns on; payment does not mean instant access.
Which terms apply?
Plan selection does not itself authorize production use. Confirm commercial or production use in a separate written agreement with Unbiased before moving live traffic.
See what your workload costs.
Prepaid usage. Account approval required.