unbiased ai

Hermes Agent, sent to Pareto

Hermes Agent was built to take any OpenAI-compatible endpoint, so give it Unbiased’s. One model block in config.yaml and every request goes to Pareto, which selects frontier and open models for each request and blends their outputs into one answer. The model changes, and so does the bill.

Choose Personal · $10/month See the setup →

Tested live with Hermes Agent 0.22.250M tokens a week

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

Set it up in config.yaml

Hermes keeps its model settings in ~/.hermes/config.yaml. Add a custom provider that points at Unbiased, or run the same changes as hermes config set commands.

  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. Add a custom provider

    In the model block, set provider to custom and base_url to https://api.unbiased.ai/v1. Hermes expects a base URL that serves /v1/chat/completions, so keep the /v1.

  3. Name the model and its context window

    Set default to pareto. Unbiased does not serve a model list, so Hermes cannot look up the context window on its own; set context_length to 1000000 yourself.

  4. Keep the key in your environment

    Set key_env to UNBIASED_API_KEY and export your key under that name, so it never lands in the config file. Then run hermes in any project.

Tested on October 20, 2026Hermes Agent 0.22.2 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.
~/.hermes/config.yaml
model:
  provider: custom
  base_url: https://api.unbiased.ai/v1
  default: pareto
  key_env: UNBIASED_API_KEY
  context_length: 1000000
Or from the shell
hermes config set model.provider custom
hermes config set model.base_url https://api.unbiased.ai/v1
hermes config set model.default pareto
hermes config set model.key_env UNBIASED_API_KEY
hermes config set model.context_length 1000000
export UNBIASED_API_KEY=your-unbiased-api-key
hermes
One-shot run
# Runs once and exits; --yolo skips approval prompts
hermes chat --oneshot --yolo -q \
  "test_chunk.py fails. Fix chunk.py and rerun it."

Prefer a wizard? Run hermes model, choose “Custom endpoint (self-hosted / VLLM / etc.)” and enter the base URL, your key and pareto. Then check that context_length is set in config.yaml.

Hermes Agent stays. The model changes.

The provider, base URL and model name change. The agent you’ve set up stays as it is.

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 Hermes Agent 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

Agents make many model calls per task, 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 Hermes Agent by the month

One API key works across Hermes Agent 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.

Hermes Agent on Pareto, questions answered

Which provider should I pick in Hermes?

Custom. In hermes model that is “Custom endpoint (self-hosted / VLLM / etc.)”; in config.yaml it is provider: custom. Hermes’s docs say OPENAI_BASE_URL only applies to the openai-api provider and that LLM_MODEL in .env has been removed, so put the base URL and model in config.yaml. We tested Hermes Agent 0.22.2 this way on October 20, 2026, over OpenAI Chat Completions: it patched a failing Python file and reran the test to green. Format compatibility is not a promise that every feature works, so tell us if one doesn’t.

Why set context_length by hand?

Hermes reads a model’s context window from the endpoint’s model list, and Unbiased does not implement GET /v1/models. Without context_length, Hermes falls back to a default far smaller than Pareto’s window. Its docs also say Hermes needs at least 64,000 tokens of context for agent work, so set the number rather than leave it to chance.

Why doesn’t /model custom find Pareto?

A bare /model custom asks the endpoint which models it has, and Unbiased has no list to give. Type the name: /model custom:pareto. /model also only switches between providers you have already set up, so add the endpoint with hermes model or config.yaml first.

Do I still need OpenRouter or a Nous Portal account?

No. With provider: custom, Hermes sends its requests to Unbiased with your Unbiased key, and your Unbiased plan pays for them. To switch back, run hermes model and pick your previous provider, or use /model in a session if that provider is still set up.

Is Pareto made by the Hermes Agent team?

No. Pareto is Unbiased’s model. Hermes Agent is made by Nous Research, and Unbiased is not affiliated with or endorsed by them. You keep Hermes Agent 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.

Send Hermes on its next errand

One model block, one key and the same terminal. Pareto starts at $10 a month for 50M tokens a week.

Choose Personal · $10/month See the setup →