Moondream: Moondream 3.1

moondream/moondream3.1-9b-a2b

Try Moondream 3.1 for free in your browser, and add it to your app for free with Puter.js AI API.

Try it free Add to your app

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

puter.ai.chat("Explain quantum computing in simple terms", {
    model: "moondream/moondream3.1-9b-a2b"
}).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: "moondream/moondream3.1-9b-a2b"
        }).then(response => {
            document.body.innerHTML = response.message.content;
        });
    </script>
</body>
</html>

Model Card

Moondream 3.1 is a vision language model from Moondream (M87 Labs). It is a sparse mixture-of-experts model with 9B total and 2B active parameters, built around visual tasks such as question answering, captioning, object detection, pointing, and segmentation.

On Moondream's preliminary benchmarks it scores 81.46 on COCO detection (Qwen3.5 27B scores 76.75), 74.61 on Dense200, and 86.01 on ChartQA (Qwen3.5 9B scores 70.72). Qwen3.5 9B scores higher on DocVQA (92.94 vs. 88.6) and GSM8K (89.46 vs. 63), and Moondream says math is not a primary target.

It accepts text and image input and returns text, with a 32,000-token context window. Tool calling is not supported. It suits developers who need detection, counting, or chart reading from images, and not general text reasoning.

Context Window 32K

tokens

Max Output 4K

tokens

Input Cost $0.3

per million tokens

Output Cost $2.5

per million tokens

Input text, image

modalities

Tool Use No

Release Date Jul 7, 2026

Try Moondream 3.1 for free

Try Moondream 3.1 instantly in your browser.
This playground uses the Puter.js AI API — no API keys or setup required.

Chat moondream/moondream3.1-9b-a2b

Chat with Moondream 3.1

More AI Models From Moondream

Find other Moondream models →

Chat

Moondream 3 Preview

Moondream 3 Preview is a vision-language model from Moondream (M87 Labs). It is a mixture-of-experts model with 9B total parameters and about 2B active per token, and it accepts text and image input. It is built for visual tasks beyond captioning. It handles object detection from descriptive queries, pointing, structured output such as JSON with minimal prompting, and OCR, though very small fonts remain a weakness. It is a hybrid reasoning model, and its reasoning steps are tied to regions of the image. The context window is 32K tokens, but the team says post-training does not yet make full use of it, which is why this is a preview. It suits developers who need low-cost image understanding, detection, or document extraction through an API.

Frequently Asked Questions

How do I use Moondream 3.1?

You can access Moondream 3.1 by Moondream 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.

Can I try Moondream 3.1 for free?

Moondream 3.1 is free to try with a Puter account. Every account includes a free AI allowance, and you can chat with it in the playground on this page. You can upgrade your account anytime for a larger allowance.

Is the Moondream 3.1 API free for developers?

Moondream 3.1 is free to integrate using the Puter.js AI API. With the User-Pays Model, you can add AI to your app for $0, since users cover their own AI usage through their Puter account.

What is the pricing for Moondream 3.1?

Moondream 3.1 costs $0.3 per 1M input tokens and $2.5 per 1M output tokens.

	Price per 1M tokens
Input	$0.3
Output	$2.5

Who created Moondream 3.1?

Moondream 3.1 was created by Moondream and released on Jul 7, 2026.

What is the context window of Moondream 3.1?

Moondream 3.1 supports a context window of 32K tokens. For reference, that is roughly equivalent to 64 pages of text.

What is the max output length of Moondream 3.1?

Moondream 3.1 can generate up to 4K tokens in a single response.

What types of input can Moondream 3.1 process?

Moondream 3.1 accepts the following input types: text, image. It produces: text.

Does Moondream 3.1 support tool use (function calling)?

No, Moondream 3.1 does not currently support tool use (function calling).

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

Yes — the Moondream 3.1 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.

Add Moondream 3.1 to your app for free

Developers can integrate Moondream 3.1 for free using the Puter.js AI API.
With the User-Pays Model, each user covers their own AI usage instead of the developer.

Get started How pricing works