Muse Wonder: benchmarks, price and rank
Muse Wonder from Meta is 10th of 123 rated AI models on Spezia's leaderboard, strongest on LMArena text and DeepSWE 1.1. It costs $4 in, $20 out per million tokens.
- 10th of 123 rated models
- Rating 1,824
- Rated on 10 benchmarks
- Closed
- Released Oct 12, 2026
- 1 million tokens of context
Where it is strongest
- 6th of 107 on LMArena text
- 2nd of 27 on DeepSWE 1.1
- 3rd of 18 on LifeSci Diamond
Where it falls behind
- 5th of 6 on OSWorld 2.1
- 9th of 17 on Terminal-Bench 4.0
- 10th of 19 on HealthBench Professional
Price: $4 in, $20 out per million tokens.
Its headline results
The benchmarks that count most in the rating, with its place among the models measured the same way, and where each number came from.
- APEX-Agents66.1%
Long tasks written by investment banking analysts, management consultants and corporate lawyers from their own work, done across files and apps in a realistic work setup.
10th of 56. Independent result from Epoch AI.
- Terminal-Bench 4.052%
Hard jobs done in a real computer terminal: building and fixing software, setting up systems, and science and engineering problems.
9th of 17. The maker's own number, from Meta, via LLM Stats.
- OSWorld 2.134.3%
The newer version of OSWorld, with longer desktop tasks that are graded at several checkpoints along the way.
5th of 6. The maker's own number, from Meta, via LLM Stats.
- Vending-Bench 2$9,432
The model runs a simulated vending machine business for a year on its own: finding suppliers, ordering stock, setting prices and paying the daily fees.
11th of 60. Independent result from Epoch AI.
- CursorBench50.4%
Ambiguous coding tasks that span many files, taken from real sessions in the Cursor code editor.
7th of 20. Independent result from Epoch AI.
- DeepSWE 1.178.8%
Software engineering tasks run in the same simple agent setup for every model.
2nd of 27. The maker's own number, from Meta, via LLM Stats.
- FrontierMath82.1%
Hundreds of new, unpublished math problems written by mathematicians, from hard undergraduate level up to research, with answers a computer can check.
18th of 76. Independent result from Epoch AI.
- LMArena text1,503
People put a question to two anonymous models and vote for the better answer.
6th of 107. Independent result from LMArena.
- LMArena WebDev1,649
People ask two anonymous models to build the same web page or app and vote for the better one.
18th of 98. Independent result from LMArena.
- HealthBench Professional60.5%
Conversations in the style of real clinicians' chats with AI, graded against rubrics written by physicians.
10th of 19. The maker's own number, from Meta, via LLM Stats.
3 more results in its full profile.
Every result with its source, what it costs per task, what jumped with its release, side by side with any other model, and answers to your questions about when to use it.
Models ranked near it
How the rating works, and every benchmark it uses: the model leaderboard.