Cursor, with Pareto in the model picker

Cursor already lets you bring your own OpenAI key and swap the address it sends to. Put your Unbiased key in that slot, override the base URL, add a custom model called pareto and pick it in chat. Your editor, rules and shortcuts stay put, and each answer comes from Pareto, which consults a blend of frontier and open models and returns one reply.

Choose Personal · $10/month See the setup →

Settings from Cursor’s docs50M tokens a week

chunk.py - shop - Cursor

chunk.py

def chunk(items, size):    """Split items into lists of length size."""    out = []    for i in range(0, len(items) - size + 1, size):    for i in range(0, len(items), size):        out.append(items[i:i + size])    return out# test_chunk.pyassert chunk([1, 2, 3, 4, 5], 2) == \    [[1, 2], [3, 4], [5]]

ChatAgent · pareto

Cursor Settings › Models

OpenAI keyyour-unbiased-api-key

Base URLhttps://api.unbiased.ai/v1

Modelpareto

Why does test_chunk.py fail?

Context chunk.py, test_chunk.py

The loop stops one group early, so the trailing 5 never gets a chunk of its own.

Suggested range(0, len(items), size)

Applied chunk.py, line 4

Terminal python3 test_chunk.py

Passing. The last slice now keeps whatever is left.

Ask about chunk.py

mainPython 3.12 · UTF-8

An illustration of Cursor on Pareto. Not a screenshot.

Set it up in Cursor Settings

Pareto goes into Cursor through its OpenAI key slot. Open Cursor Settings and go to Models.

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.
Add it as your OpenAI key
Under API Keys, paste your Unbiased key into OpenAI API Key and turn the Use OpenAI API key toggle on. Pasting the key and turning on the toggle are two separate steps.
Override the base URL
Turn on Override OpenAI Base URL and enter https://api.unbiased.ai/v1. Keep the /v1: Cursor appends OpenAI paths such as /chat/completions and, in Agent, /responses to this address. Unbiased serves both.
Add pareto as a custom model
Add a custom model named pareto and switch it on. Unbiased does not publish a model list, so Cursor cannot fetch the name for you; type it by hand.
Pick it and start a new chat
Open a new chat, choose pareto in the model picker under the input box and send a message. If you turned the key on in the middle of a chat, start a new one so the change takes effect.

Settings follow Cursor’s own documentation: Cursor Help: Bring your own API key.

Cursor Settings › Models

# API Keys
OpenAI API Key            your-unbiased-api-key
Use OpenAI API key        on
Override OpenAI Base URL  https://api.unbiased.ai/v1

# Models › Add Custom Model
pareto

Check the key outside Cursor

curl https://api.unbiased.ai/v1/chat/completions \
  -H "Authorization: Bearer your-unbiased-api-key" \
  -H "Content-Type: application/json" \
  -d '{"model":"pareto","messages":[{"role":"user","content":"Say ok"}]}'

While the key and override are on, Cursor sends every model except Claude and Gemini to this address, including its own GPT, Composer and Grok models. Turn the OpenAI key off in Cursor Settings › Models when you want those back. Cursor’s help page covers the key field; Cursor staff describe the base URL override and turning it off, which models it covers and the Use OpenAI API key toggle on the Cursor forum.

Cursor stays. The chat model changes.

Your own key covers Cursor’s chat models. Everything around the chat is the editor you already set up.

Extensions, keybindings and project rulesUnchanged
Tab completionCursor’s built-in models
Chat modelpareto, a custom model
Where chat requests goThrough Cursor to api.unbiased.ai
Who bills the chat tokensYour Unbiased plan, from $10 a month

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 Cursor 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
DeepSWE v1.1Changes an agent makes in a real codebase: $0.24/task

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 the chat model by the month

Cursor bills for Cursor. Your Unbiased plan covers the Pareto tokens, and the same key works in any other tool you point at Pareto.

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.

Cursor on Pareto, questions answered

Does Pareto power Tab completion in Cursor?

No. Cursor’s help page says your own API keys work only with chat models, and Tab completion keeps using Cursor’s built-in models.

Does Agent mode work with Pareto?

Cursor sends custom models in OpenAI formats (Chat Completions, and Responses in Agent), and Unbiased serves both. Cursor’s forum has a report of a custom model running in a new Agent chat once the key toggle was on, alongside a plan check that Cursor staff say should not apply and are tracking. Format compatibility is not a promise that every Cursor feature works. If you see a bad model name error, check that Use OpenAI API key is on and start a new chat. If it persists, Cursor staff suggest deleting the key, adding it again, then flipping the toggle off and on once.

Why do Cursor’s own models stop working when the override is on?

Cursor staff say the OpenAI key and base URL override apply to every model that is not a Claude or Gemini model, so Cursor’s own GPT, Composer and Grok models are also sent to api.unbiased.ai, which does not serve them. Cursor staff suggest turning Override OpenAI Base URL off when you want Cursor’s models and back on for pareto.

Do I still need a Cursor subscription?

You still sign in to Cursor. On individual plans, requests made with your own key do not draw from Cursor’s included usage. On Teams and Enterprise, Cursor adds its own token rate to own-key requests. Your Unbiased plan pays for the Pareto tokens.

Does my key go to Cursor?

Yes. Cursor’s docs say every request goes through Cursor’s servers for final prompt building, so your key travels with each request and the base URL must be reachable from the internet, which api.unbiased.ai is. Cursor says the key is not stored after the request. Cursor’s Zero Data Retention policy does not cover own-key requests.

Is Pareto made by the Cursor team?

No. Pareto is Unbiased’s model. Cursor is made by Anysphere, and Unbiased is not affiliated with or endorsed by them. You keep Cursor 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.

Add one name to Cursor’s model picker

A key, a base URL and the word pareto. Plans start at $10 a month for 50M tokens a week.