Early-bird pricing (ends Dec 25), locked in while you stay subscribed: Go $10/mo (normally $20), Plus with Opus 5.5 just $30/mo (normally $50)Subscribe →

Muse Spark 1.3 Contributor

Released 2 September 2026. This model id uses the Muse Spark 1.3 weights and a 1,048,576-token context. The lower rate comes with a data term: prompts and completions may be used to improve Meta products.

Same Muse Spark 1.3 weights

Context length, multimodal input, and tool calling match Muse Spark 1.3. The model id on this page is muse-spark-1.3-contributor.

Rate on this page

Input $0.10, output $0.20, and cached input $0.002 per million tokens. The figures at the top of the page are the billed rates.

Text, image, PDF, video, and audio

The API accepts those inputs and returns text. Send them in the Chat Completions message body.

Coding and agent work

Public repositories, documentation, prototypes, and agent loops where the prompt can be shared under the term below.

High-Leverage Engineering & Research Scenarios

Designed to maximize iteration speed for developers, researchers, and innovative agent builders.

Continuous Agent Benchmarking & Loops

Run hundreds of autonomous tool-calling iterations, evaluation matrices, and multi-agent test environments without budget friction.

Open-Source Engineering & Documentation

Analyze open repositories, author comprehensive documentation, generate regression test matrices, and refactor open modules at scale.

High-Volume PDF & Technical Ingestion

Extract structured intelligence from thousands of public PDF manuals, technical papers, visual schematics, and media archives.

Synthetic Data Generation & Rapid Prototyping

Synthesize rich instruction-tuning datasets, generate interactive learning materials, and power community developer tools affordably.

Data use

Read this before sending customer code, credentials, or internal documents.

Prompts may be used by Meta

On muse-spark-1.3-contributor, prompts and completions may be used to improve Meta products.

What to send instead

Client code, credentials, production logs, and internal documents belong on MiMo V2.6 Flash or MiMo V2.6 Pro.

How to call it

POST /api/v1/chat/completions with model set to muse-spark-1.3-contributor. Tool calls and JSON output use the same Chat Completions fields.

Cached input

Tokens reported as cache reads are billed at $0.002 per million. Other input tokens use the input rate.

Quick Integration via Standard Chat Completions

Call Muse Spark 1.3 Contributor using standard OpenAI-compatible libraries.

  1. Specify Contributor Model Identifier

    Pass muse-spark-1.3-contributor in your Chat Completions payload.

  2. Submit Multimodal Reasoning Request

    POST https://aireiter.com/api/v1/chat/completions Authorization: Bearer $AIREITER_API_KEY Content-Type: application/json { "model": "muse-spark-1.3-contributor", "messages": [ {"role": "user", "content": "Analyze this algorithmic workflow and propose structural optimizations."} ], "temperature": 0.3 }

  3. Check the response

    Read the message content, finish reason, and token usage. Retry a timeout or a 502 with a bounded backoff.

Muse Spark 1.3 Contributor FAQ

Frequently asked questions regarding capabilities, pricing, and data policy.

01Will Meta use what I send?

Yes. Prompts and completions sent to muse-spark-1.3-contributor may be used to improve Meta products. Send client code, credentials, and internal documents to MiMo V2.6 Flash or MiMo V2.6 Pro instead.

02Is this the same model as Muse Spark 1.3?

It uses the Muse Spark 1.3 weights. The model id for this page is muse-spark-1.3-contributor.

03What can the API accept?

Text, images, PDFs, video, and audio. The response is text.

04What model id do I send?

muse-spark-1.3-contributor.

05How long is the context window?

1,048,576 tokens.