Muse Spark 1.2: benchmark results
Meta's Muse Spark 1.2 model, a coding-focused update to Muse Spark 1.1 released alongside the Muse Code terminal agent; API-only, with no disclosed parameter count. Provider: Meta. Released 2026-08-05. Access: API.
Aggregate Rating 1799 ± 7, rank #69 of 1635 rated models, from 413 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Chatbot Arena (Text - French) | 1534 | Arena Score | 100 |
| Chatbot Arena (Text - Legal & Government) | 1535 | Arena Score | 100 |
| Chatbot Arena (Text - Spanish) | 1505 | Arena Score | 100 |
| MuSP-Bench - abc - Composer/title - 48 Questions | 54.17 | Accuracy (%) | 100 |
| MuSP-Bench - abc+audio - S&P - 72 Questions | 6.94 | Accuracy (%) | 100 |
| MuSP-Bench - midi-numeric - Composer/title - 48 Questions | 22.92 | Accuracy (%) | 100 |
| MuSP-Bench - midi-numeric - Tonality/style - 48 Questions | 66.67 | Accuracy (%) | 100 |
| MuSP-Bench - pdf - Composer/title - 48 Questions | 79.17 | Accuracy (%) | 100 |
| MuSP-Bench - pdf - Tonality/style - 48 Questions | 91.67 | Accuracy (%) | 100 |
| MuSP-Bench - pdf+audio - S&P - 72 Questions | 6.94 | Accuracy (%) | 100 |
| Qwen3.8 Omni Report - OmniCloze | 65.3 | Score (%) | 100 |
| RxScribe Bench - Fabrications per Run | 2.28 | Fabrication events per run (lower is better; invented medici | 100 |
Interactive version: theaggregate.ai/model?slug=muse-spark-1-2 · How It Works · Data refreshed daily, snapshot 2026-11-05.