unbiased ai

Kilo Code, with Pareto as the custom provider

Kilo Code already lets you bring your own provider. Add Unbiased as an OpenAI Compatible provider, type in one model called pareto, and Kilo sends its requests to Pareto, which selects frontier and open models and blends their outputs into one answer. Your modes, rules and approvals stay as they were.

Choose Personal · $10/month See the setup →

Tested live with Kilo Code CLI 7.8.1550M tokens a week

A demonstration of Kilo Code on Pareto, recreated from a real Kilo Code CLI run. Not a screenshot.

Add Unbiased as a custom provider

Kilo Code takes any OpenAI Compatible endpoint as a custom provider. The CLI reads it from ~/.config/kilo/kilo.json; the VS Code extension adds it from the Providers tab. Both want the same four facts: a base URL, a key, a provider ID and the model pareto.

  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. In the editor: open Providers

    Click the gear icon in the Kilo Code panel, go to the Providers tab and click Custom provider at the bottom of the list. Enter unbiased as the Provider ID, Unbiased as the Display name, OpenAI Compatible as the Provider API, https://api.unbiased.ai/v1 as the Base URL and your key, then add pareto under Models by hand and click Submit. These steps follow Kilo’s Custom Models guide.

  3. In the terminal: write the global config

    Put the config block in ~/.config/kilo/kilo.json and export UNBIASED_API_KEY. Keep it in the global file: a project-level kilo.json ignores {env:...} references.

  4. Pick unbiased/pareto

    Choose Pareto in the model picker, or pass -m unbiased/pareto to kilo. Kilo names models as provider ID, then model ID. Restart the CLI after editing the config.

Tested on October 20, 2026Kilo Code CLI 7.8.15 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.
~/.config/kilo/kilo.json
{
  "$schema": "https://app.kilo.ai/config.json",
  "model": "unbiased/pareto",
  "provider": {
    "unbiased": {
      "name": "Unbiased",
      "npm": "@ai-sdk/openai-compatible",
      "options": {
        "baseURL": "https://api.unbiased.ai/v1",
        "apiKey": "{env:UNBIASED_API_KEY}"
      },
      "models": {
        "pareto": {
          "name": "Pareto",
          "tool_call": true,
          "limit": { "context": 1000000, "output": 131072 }
        }
      }
    }
  }
}
Shell
export UNBIASED_API_KEY=your-unbiased-api-key
# --auto follows your auto-approval settings
kilo run --auto -m unbiased/pareto "test_chunk.py fails. Fix the bug in chunk.py."

Keep limit.context and limit.output set. Kilo’s docs warn that a context of 0 turns off compaction, so a long session grows until the provider rejects it.

Kilo Code stays. The model changes.

One provider entry and one model name are new. The modes, rules and approvals you set up are not.

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 Kilo Code 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 Kilo Code by the month

One API key works across Kilo Code 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.

Kilo Code on Pareto, questions answered

Which Provider API should I choose in Kilo Code?

Choose OpenAI Compatible and use the base URL https://api.unbiased.ai/v1, with the /v1 on the end. Kilo then sends Chat Completions requests, which is the format Unbiased serves there.

Why doesn’t Fetch models find Pareto?

Fetch models asks the provider for its model list, and Unbiased does not publish one. Type pareto into Models yourself, or declare it under models in kilo.json, and it shows up in the picker. Mark it tool_call: true in the config so Kilo offers the agent its edit and shell tools.

Why is my key ignored in a project kilo.json?

Kilo resolves {env:VAR} only in trusted config: the global ~/.config/kilo folder, or KILO_CONFIG and KILO_CONFIG_CONTENT. A kilo.json committed to a repository ignores it and logs a warning. Keep the provider and key in your global config.

Do I need Kilo credits or a Kilo account for this?

Not for the model calls. Requests to a custom provider go to that provider with your key, and your Unbiased plan pays for them. To switch back, pick another model in the picker or drop the -m flag.

Has Unbiased tested Kilo Code with Pareto?

Yes, in the terminal. We ran Kilo Code CLI 7.8.15 against Pareto on October 20, 2026: it read a failing Python test, fixed the bug in chunk.py and reran the test to green. The editor settings above follow Kilo’s documentation. Format compatibility is not a promise that every feature works, so tell us if one doesn’t.

Is Pareto made by the Kilo Code team?

No. Pareto is Unbiased’s model. Kilo Code is made by Kilo Code (Kilo-Org), and Unbiased is not affiliated with or endorsed by them. You keep Kilo Code 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 Kilo the failing test and a new model

One custom provider, one key and the same agent in your editor or terminal. Pareto starts at $10 a month for 50M tokens a week.

Choose Personal · $10/month See the setup →