BU Ultrafast - Browser Use

BU Ultrafast runs Browser Use Pi in ultrafast mode on GPT-6 Luna with low reasoning and OpenAI’s priority tier. It is available on every plan, including free.

Pick one

Task	`model`
Short, interaction-heavy flows	`bu-ultrafast`
The same kind of flow with more steps, where `bu-ultrafast` misses	`bu-fast` (medium reasoning, standard tier)
Research, bulk collection, files, long instructions	Omit `model` for the default, GPT-6 Luna with High reasoning

from browser_use_sdk.v4 import BrowserUse

with BrowserUse() as client:
    run = client.runs.create(
        "Find the top Hacker News story",
        model="bu-ultrafast",
    )
    result = client.runs.wait_for_completion(run.id)
    print(result.result)

import { BrowserUse } from "browser-use-sdk/v4";

const client = new BrowserUse();
const run = await client.runs.create({
  task: "Find the top Hacker News story",
  model: "bu-ultrafast",
});
const result = await client.runs.waitForCompletion(run.id);
console.log(result.result);

curl https://api.browser-use.com/api/v4/runs \
  -H "X-Browser-Use-API-Key: $BROWSER_USE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"task":"Find the top Hacker News story","model":"bu-ultrafast"}'

Limits

bu-ultrafast and bu-fast are presets and require Browser Use Pi. They take no modelParams, mode or harness: "browsercode"; sending any of them returns 422.
The preset sticks to the session. Follow-ups keep it; a follow-up that sets another model or modelParams returns 422.
New sessions run without the managed proxy unless you set browserSettings.proxyCountryCode (for example "us"). See Proxies.
HIPAA and Bedrock-only projects run the preset on the standard tier.

Pricing

Both presets bill GPT-6 Luna tokens. bu-ultrafast uses OpenAI’s priority tier, billed at the Fast-mode rate (2x Standard, see Thinking levels); bu-fast bills at the Standard rate. With your own OpenAI key, OpenAI bills the tokens to your account, plus the 0.2x Browser Use orchestration fee.

⌘I