Unbiased API
Access Unbiased instantly with Puter.js, and add AI to any app in a few lines of code without backend or API keys.
// npm install @heyputer/puter.js
import { puter } from '@heyputer/puter.js';
puter.ai.chat("Explain AI like I'm five!", {
model: "unbiased/pareto"
}).then(response => {
console.log(response);
});
<html>
<body>
<script src="https://js.puter.com/v2/"></script>
<script>
puter.ai.chat("Explain AI like I'm five!", {
model: "unbiased/pareto"
}).then(response => {
console.log(response);
});
</script>
</body>
</html>
List of Unbiased Models
Pareto 26.9
unbiased/pareto-26.9
Pareto 26.9 is a blended AI model from Unbiased AI, the platform of Circuit & Chisel, a remote-first team in the US and Canada. It runs several frontier and open models on each request and returns one answer. It does not switch models within a conversation, so prompt caching still applies. On Unbiased's published scores, Pareto 26.9 gets 74 on DeepSWE, 88 on ArXivMath, and 78 on MMMU-Pro. It scores 51 on Terminal-Bench 4.0 and 49 on Humanity's Last Exam without tools, which are below the frontier models. Because the underlying model mix can change between releases, scores cannot be attributed to a single model. It supports tool calling and has a 262,144-token context window with up to 131,072 output tokens. It suits coding and agent workflows that need consistent behavior across a session.
ChatPareto 26.10 Preview
unbiased/pareto-26.10-preview
Pareto 26.10 Preview is a composite AI model from Unbiased AI, built by Circuit & Chisel, a remote-first team in the US and Canada. It runs several models on the same request and returns one answer, and it is aimed at research, coding, and agentic workflows. On Unbiased's preliminary, vendor-run benchmarks it scores 69.9% on DeepSWE v1.1 at $0.24 per task, 50.8% on Terminal-Bench 4.0, 49.9% on Humanity's Last Exam (text only), and 92.4% on GPQA-Diamond. On DeepSWE it is level with Fable 5 (70.0%, $13.50 per task). On Terminal-Bench, GPT 6 Astra (58) and Claude Fable 5.1 (56) score higher. It accepts text and image input, supports tool calling, and has a 1,048,576-token context window with up to 131,072 output tokens. It suits engineering teams, platform teams, and agent builders who want a low cost per task.
ChatPareto
unbiased/pareto
Pareto is a blended AI model from Unbiased AI, built by Circuit & Chisel, a remote-first team based in the US and Canada. Rather than routing a request to one model, it runs several models against the same request and returns the best answer, without switching mid-conversation, which preserves prompt caching. On Unbiased's published benchmarks, Pareto 26.9 scores 88/100 on ArXivMath and 78/100 on MMMU-Pro, and ties GPT 6 Astra and DeepSeek 4.1 Flash at 74/100 on the DeepSWE coding benchmark. It trails Fable 5.1 and GPT 6 Astra on Terminal-Bench 4.0 and HLE, and Unbiased publishes those losses alongside its wins rather than only citing favorable results. It accepts text and image input and supports tool calling. It targets engineering teams evaluating model performance, platform teams looking for a cost-effective general-purpose option, and developers building agents that need consistent responses across a session.
Frequently Asked Questions
The Unbiased API gives you access to models for AI chat. Through Puter.js, you can start using Unbiased models instantly with zero setup or configuration.
Puter.js supports a variety of Unbiased models, including Pareto 26.9, Pareto 26.10 Preview, and Pareto. Find all AI models supported by Puter.js in the AI model list.
With the User-Pays model, users cover their own AI costs through their Puter account. This means you can build apps without worrying about infrastructure expenses.
Puter.js is a JavaScript library that provides access to AI, storage, and other cloud services directly from a single API. It handles authentication, infrastructure, and scaling so you can focus on building your app.
Yes — the Unbiased API through Puter.js works with any JavaScript framework, Node.js, or plain HTML. Just include the library and start building. See the documentation for more details.