Muse Wonder (Low): benchmark results
Provider: Meta. Released 2026-10-12. Access: API.
Aggregate Rating 1849 ± 26, rank #86 of 2201 rated models, from 17 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AI for Education Pedagogy - Maths | 92.06 | Accuracy (%) | 90.3 |
| AI for Education Pedagogy | 89.66 | Accuracy (%) | 90.1 |
| LisanBench | 0.24 | Mean Path Length / Current Maximum | 88.6 |
| AI for Education Pedagogy - Secondary | 88.84 | Accuracy (%) | 88.5 |
| AI for Education Pedagogy - Science | 92.35 | Accuracy (%) | 86.7 |
| AI for Education SEND | 83.94 | Accuracy (%) | 84.5 |
| AI for Education Pedagogy - Primary | 92.02 | Accuracy (%) | 80.2 |
| AI for Education Pedagogy - Technology | 83.96 | Accuracy (%) | 75.4 |
| AI for Education Pedagogy - Social studies | 84.55 | Accuracy (%) | 71.9 |
| cua-speedrun - OSWorld 2.0 (52 Tasks) | 67.86 | Mean verifier score (%), partial credit over the task set | 63 |
| EnigmaForge - Fact F1 | 87.9 | World-fact recovery F1 (%) over 600 instances, all items | 59.4 |
| cua-speedrun - OSWorld (50 Tasks) | 83.62 | Mean verifier score (%), partial credit over the task set | 56.2 |
Interactive version: theaggregate.ai/model?slug=muse-wonder-low · How It Works · Data refreshed daily, snapshot 2026-11-12.