Prime Intellect

Prime Intellect: INTELLECT-3

prime-intellect/intellect-3

Access INTELLECT-3 from Prime Intellect using Puter.js AI API.

Get Started 
// npm install @heyputer/puter.js
import { puter } from '@heyputer/puter.js';

puter.ai.chat("Explain quantum computing in simple terms", {
    model: "prime-intellect/intellect-3"
}).then(response => {
    document.body.innerHTML = response.message.content;
});

<html>
<body>
    <script src="https://js.puter.com/v2/"></script>
    <script>
        puter.ai.chat("Explain quantum computing in simple terms", {
            model: "prime-intellect/intellect-3"
        }).then(response => {
            document.body.innerHTML = response.message.content;
        });
    </script>
</body>
</html>

# pip install openai
from openai import OpenAI

client = OpenAI(
    base_url="https://api.puter.com/puterai/openai/v1/",
    api_key="YOUR_PUTER_AUTH_TOKEN",
)

response = client.chat.completions.create(
    model="prime-intellect/intellect-3",
    messages=[
        {"role": "user", "content": "Explain quantum computing in simple terms"}
    ],
)

print(response.choices[0].message.content)

curl https://api.puter.com/puterai/openai/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_PUTER_AUTH_TOKEN" \
  -d '{
    "model": "prime-intellect/intellect-3",
    "messages": [
      {"role": "user", "content": "Explain quantum computing in simple terms"}
    ]
  }'

Model Card

INTELLECT-3 is a 106B-parameter Mixture-of-Experts reasoning model from Prime Intellect, with 12B active parameters per forward pass. It was post-trained from GLM-4.5-Air-Base using supervised fine-tuning followed by large-scale reinforcement learning.

The model excels at math, code, science, and multi-step reasoning tasks. It scores 98.1% on MATH-500, 90.8% on AIME 2024, 69.3% on LiveCodeBench v6, and 74.4% on GPQA Diamond — outperforming the base GLM-4.5-Air it was trained from and competing with larger frontier models.

Its MoE architecture keeps inference efficient despite the large total parameter count, making it a strong choice for developers who need high reasoning performance without the cost profile of much larger dense models. Fully open-weight under the MIT license, with a 131K token context window.

Context Window 131K

tokens

Max Output 131K

tokens

Input Cost $0.2

per million tokens

Output Cost $1.1

per million tokens

Release Date May 28, 2025

 

Model Playground

Try INTELLECT-3 instantly in your browser.
This playground uses the Puter.js AI API — no API keys or setup required.

Chat prime-intellect/intellect-3
Prime Intellect
Chat with INTELLECT-3
Powered by Puter.js

Benchmarks

How INTELLECT-3 performs on standard evaluations.

Artificial Analysis
Intelligence Index
22.2
Better than 58% of tracked models
Artificial Analysis
Coding Index
19.1
Better than 53% of tracked models
Artificial Analysis
Math Index
88.0
Better than 85% of tracked models
BenchmarkScore
GPQA Diamond Graduate-level science Q&A
76.1%
Humanity's Last Exam Cross-domain reasoning
12.1%
LiveCodeBench Recent coding problems
77.7%
SciCode Scientific programming
39.1%
AIME 2025 Advanced math exam
88.0%
IFBench Instruction following
34.0%
LCR Long-context reasoning
32.3%
Terminal-Bench Hard Agentic terminal tasks
9.1%
τ²-Bench Tool use / agents
26.6%

Scores sourced from Artificial Analysis.

Frequently Asked Questions

How do I use INTELLECT-3?

You can access INTELLECT-3 by Prime Intellect through Puter.js AI API. Include the library in your web app or Node.js project and start making calls with just a few lines of JavaScript — no backend and no configuration required. You can also use it with Python or cURL via Puter's OpenAI-compatible API.

Is INTELLECT-3 free?

Yes, it is free if you're using it through Puter.js. With the User-Pays Model, you can add INTELLECT-3 to your app at no cost — your users pay for their own AI usage directly, making it completely free for you as a developer.

What is the pricing for INTELLECT-3?
Pricing for INTELLECT-3 is based on the number of input and output tokens used per request.
Price per 1M tokens
Input$0.2
Output$1.1
Who created INTELLECT-3?

INTELLECT-3 was created by Prime Intellect and released on May 28, 2025.

What is the context window of INTELLECT-3?

INTELLECT-3 supports a context window of 131K tokens. For reference, that is roughly equivalent to 262 pages of text.

What is the max output length of INTELLECT-3?

INTELLECT-3 can generate up to 131K tokens in a single response.

Does it work with React / Vue / Vanilla JS / Node / etc.?

Yes — the INTELLECT-3 API works with any JavaScript framework, Node.js, or plain HTML through Puter.js. Just include the library and start building. See the documentation for more details.

Get started with Puter.js

Add INTELLECT-3 to your app without worrying about API keys or setup.

Read the Docs View Tutorials