Meta · Released Oct 12, 2026Muse Wonder
Muse Wonder is a frontier multimodal reasoning model from Meta for personal and professional agents, long-horizon coding, and computer use. It combines Active Memory, which carries relevant context across conversations, with Active Focus, which keeps it working through blockers toward a goal.
- Capability
- 161.7ECI · #13 of 165
- Input
- $4.00per 1M tokens
- Output
- $20.00per 1M tokens
- Context
- 1.05M131K max output
What it costs
- Input
- $4.00per million tokens
- Output
- $20.00per million tokens
- Cached input
- $0.20per million tokens
- Blended (3:1)
- $8.00pricier than 93% of priced models
Typical monthly bills
| Monthly usage | Estimated cost |
|---|---|
| Side project2M input + 0.5M output tokens | $18.00 |
| Team assistant25M input + 5M output tokens | $200.00 |
| Production app250M input + 50M output tokens | $2,000 |
Official Meta API price, as listed on models.dev.
Independent benchmarks
#13 of 165 scored models
The shaded band is Epoch AI’s confidence range (159.2–164.6); the tick marks the median scored model.
Scores from Epoch AI, run independently of Meta.
The details
- Lab
- Meta
- Released
- Oct 12, 2026
- Context window
- 1,048,576 tokens
- Max output
- 131,072 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-wonder-1.0on Meta- Availability
- 9 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 |
|---|---|---|---|
| Harvey Legal Agent Benchmark v1.1 | 10 | max effort; Artificial Analysis LAB-AA | Source |
| Finance Agent vv2 | 56.6 | max effort | Source |
| HealthBench Professional | 60.5 | max effort | Source |
| AutomationBench | 39.1 | max effort; Artificial Analysis AutomationBench-AA | Source |
| GDP.pdf | 30.8 | max effort | Source |
| JobBench | 70.5 mean reward | max effort | Source |
| OSWorld v2.1 | 72.6 | max effort | Source |
| OSWorld v2.1 | 34.3 binary completion rate | max effort | Source |
| Terminal-Bench v4.0 | 52 | max effort; Artificial Analysis Terminal-Bench 4.0-AA | Source |
| DeepSWE v1.1 | 78.8 | max effort | Source |
| SWE-Marathon | 53.1 | max effort | Source |
| ProgramBench | 50 | max effort | Source |
More from Meta
About Muse Wonder
How much does Muse Wonder cost?
Muse Wonder costs $4.00 input / $20.00 output per million tokens (official Meta API price). Cached input is $0.20 per million tokens. At a 3:1 input-to-output mix that is $8.00 per million tokens, more expensive than 93% of the 445 priced models we track.
What is the context window of Muse Wonder?
Muse Wonder accepts up to 1,048,576 tokens per request and can write up to 131,072 tokens in one response.
How good is Muse Wonder?
Epoch AI gives Muse Wonder a Capabilities Index score of 161.7 (likely range 159.2–164.6), ranking it #13 of 165 models Epoch has scored. Epoch AI benchmark results: FrontierMath Tiers 1–3 82.1%, OTIS Mock AIME 2024–2025 99.4%.
Is Muse Wonder open source?
No. Muse Wonder is proprietary; you use it through Meta’s API or partner platforms.
What inputs does Muse Wonder support?
Muse Wonder 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 Wonder released?
Meta released Muse Wonder on Oct 12, 2026.
What are the best alternatives to Muse Wonder?
The closest current models from other labs on capability, price and release date are GPT-5.6 Sol (OpenAI, ECI 162.6, $4.00 / $20.00), Claude Opus 5 (Anthropic, ECI 164.5, $5.00 / $25.00), Kimi K3 (Moonshot AI, ECI 158.3, $3.00 / $15.00) and Gemini 4 Argon (Google, ECI 166.1, $2.00 / $10.00).