# Meta Model API — AI Compare Ninja

> Meta's developer API for Muse Wonder (preview, 2026-10-12) and Muse Spark 1.3, billed per token.

## Overview
- **Category:** Development & Code (Agent & RAG frameworks)
- **Status:** Active · checked - Active. Developer API for Muse Wonder (preview, 2026-10-12) and Muse Spark (currently 1.3), plus Muse Image and Muse Voice Transcribe. Muse Wonder Full is not available on this API. Separate from the Meta AI consumer assistant and the Muse Code CLI.
- **Paid from:** $1.25 / $4.25 / per 1M input/output tokens (Standard)
- **Free access:** not published
- **Website:** https://dev.meta.ai/docs/models/
- **Canonical URL:** https://www.aicompare.ninja/en/tool/meta-model-api

## What it does
- Muse Wonder (preview, 2026-10-12) for agentic tasks and coding, Standard mode only
- Muse Spark 1.3 remains the lower-priced standard model for agentic coding
- Standard $1.25/$4.25; Contributor $0.10/$0.20 per 1M input/output tokens
- Muse Wonder: input $4, cached input $0.20, output $20 per 1M tokens; web search $2.50 per 1,000 queries
- OpenAI- and Anthropic-compatible clients supported
- 1,048,576-token context; 3,000 requests per minute and 4,000,000 tokens per minute
- Also Muse Image and Muse Voice Transcribe on the same platform

## In brief
Meta Model API is the developer surface for Muse Wonder, Muse Spark, Muse Image, and Muse Voice Transcribe. Muse Wonder (preview, 2026-10-12, model muse-wonder-1.0) is for personal and professional agentic tasks and coding, with Active Memory and Active Focus. Input is text, image, video, audio, and PDF. Output is text. The context is 1,048,576 tokens. Reasoning effort runs from minimal to max, and max does not select Muse Wonder Full. Muse Spark 1.3 remains the lower-priced model for agentic and coding workloads.

Muse Wonder on this API is Standard mode only: input $4, cached input $0.20, and output $20 per 1M tokens. Contributor mode is not available for Muse Wonder here. Web search costs $2.50 per 1,000 queries, on top of the tokens. The limit is 3,000 requests per minute and 4,000,000 tokens per minute. Muse Spark 1.3 Standard stays at input $1.25, cached input $0.15, and output $4.25 per 1M tokens. Muse Spark Contributor is input $0.10, cached input $0.002, and output $0.20 per 1M tokens, and Meta may use prompts and completions for training. Muse Wonder Full is not available on this API, and its token price is not published. Separate from the Meta AI consumer product.


## Limits
Standalone developer API on dev.meta.ai, separate from Meta AI (consumer) and Muse Code (CLI). Full price table on the pricing page. Muse Wonder (preview, 2026-10-12) runs here in Standard mode only. Muse Wonder Full is not available on this API.

## Pricing tiers
| Tier | Price | Period | What you get |
| --- | --- | --- | --- |
| Muse Wonder | $4 / $0.20 / $20 | per 1M input/output tokens | Model muse-wonder-1.0, preview, released 2026-10-12; Input $4, cached input $0.20, output $20 per 1M tokens; Standard mode only; Contributor mode is not available on this API; Input text, image, video, audio, and PDF; output text; 1,048,576-token context; reasoning effort from minimal to max, and max does not select Full; Web search $2.50 per 1,000 queries, on top of the tokens; Rate limit 3,000 requests per minute and 4,000,000 tokens per minute; Muse Wonder Full is not available on this API; token price not published |
| Standard | $1.25 / $4.25 | per 1M input/output tokens | Models muse-spark-1.3, 1.2, 1.1; Input $1.25, cache read $0.15, output $4.25 per 1M tokens; Data not used for training; 1,048,576-token context; Base URL https://api.meta.ai/v1 |
| Contributor | $0.10 / $0.20 | per 1M input/output tokens | Models muse-spark-1.3-contributor and 1.2-contributor; Input $0.10, cache read $0.002, output $0.20 per 1M tokens; Prompts and completions may be used for training; Lower rate limits than Standard; Does not include Muse Wonder. Contributor mode is not available for Muse Wonder on this API. |

## Sources
All details regarding plans, limits, and features are based on official vendor documentation and the primary sources verified below.

- https://dev.meta.ai/docs/models/
- https://dev.meta.ai/docs/pricing-rate-limits
- https://research.meta.ai/blog/muse-wonder
- https://research.meta.ai/static/muse-wonder-model-card

---
*Source: AI Compare Ninja (2026) - https://www.aicompare.ninja/en/tool/meta-model-api*