Muse Ember
Muse Ember is Meta's smallest and most efficient reasoning model, built for the work agents repeat all day: tool calls, coding subtasks, routing, extraction, and long-document parsing. It pairs with Muse Wonder or Muse Spark, which plan while Muse Ember handles the volume.
- Capability
- —No ECI score yet
- Input
- $0.10per 1M tokens
- Output
- $0.50per 1M tokens
- Context
- 1.05M524K max output
What it costs
- Input
- $0.10per million tokens
- Output
- $0.50per million tokens
- Cached input
- $0.004per million tokens
- Over 200K tokens
- $0.30 / $1.50input / output per million
- Blended (3:1)
- $0.20cheaper than 75% of priced models
Typical monthly bills
| Monthly usage | Estimated cost |
|---|---|
| Side project2M input + 0.5M output tokens | $0.45 |
| Team assistant25M input + 5M output tokens | $5.00 |
| Production app250M input + 50M output tokens | $50.00 |
Official Meta API price, as listed on models.dev.
Independent benchmarks
Epoch AI has not published a Capabilities Index score for Muse Ember yet.
The details
- Lab
- Meta
- Released
- Nov 9, 2026
- Context window
- 1,048,576 tokens
- Max output
- 524,288 tokens
- Inputs
- Text, Images, PDFs, Audio, Video
- Output
- Text
- Reasoning
- Adjustable effort minimal · low · medium · high · xhigh · max
- Tool calling
- Yes
- Structured output
- Yes
- Weights
- Proprietary
- API model ID
muse-ember-1.0on Meta- Availability
- 6 API providerslisted on models.dev
What Meta claims
Published by the lab at launch. Settings vary, so compare these only with care.
| Benchmark | Score | Setting | Source |
|---|---|---|---|
| AutomationBench | 37.8 | max effort; Artificial Analysis AutomationBench-AA | Source |
| Agents' Last Exam | 61.6 | max effort | Source |
| Terminal-Bench v4.0 | 40.9 | max effort | Source |
| FrontierCode v2.0 | 27.2 | max effort; Artificial Analysis FrontierCode-AA | Source |
| GDP.pdf | 22.8 | max effort; Artificial Analysis GDP.pdf-AA | Source |
| OSWorld v2.1 | 69.5 | max effort | Source |
| OSWorld v2.1 | 33.8 binary completion rate | max effort | Source |
| PerceptionBench | 80.6 | max effort; with tools | Source |
| Artificial Analysis Intelligence Index v5 | 36 | max effort | Source |
More from Meta
About Muse Ember
How much does Muse Ember cost?
Muse Ember costs $0.10 input / $0.50 output per million tokens (official Meta API price). Cached input is $0.004 per million tokens. Requests over 200K tokens are billed at $0.30 input / $1.50 output. At a 3:1 input-to-output mix that is $0.20 per million tokens, cheaper than 75% of the 469 priced models we track.
What is the context window of Muse Ember?
Muse Ember accepts up to 1,048,576 tokens per request and can write up to 524,288 tokens in one response.
How good is Muse Ember?
Epoch AI has not published a Capabilities Index score for Muse Ember yet.
Is Muse Ember open source?
No. Muse Ember is proprietary; you use it through Meta’s API or partner platforms.
What inputs does Muse Ember support?
Muse Ember accepts text, images, PDFs, audio and video and replies in text. It is a reasoning model with minimal, low, medium, high, xhigh and max effort settings, supports tool calling and can return structured JSON output.
When was Muse Ember released?
Meta released Muse Ember on Nov 9, 2026.
What are the best alternatives to Muse Ember?
The closest current models from other labs on capability, price and release date are Viv Fast (Vivgrid, $0.13 / $0.40), Mercury 3 (Inception, $0.10 / $0.35), MiMo-V2.6-Flash (Xiaomi, $0.14 / $0.28) and Solar Mini 4 (Upstage, $0.10 / $0.40).