MuSP-Bench - pdf - Composer/title - 48 Questions: leaderboard
Frozen released MuSP experiment: 48 questions, pdf input, Composer/title subset; deterministic typed-answer scoring. Two saved historical-prompt responses remain in the published main table and are marked upstream; adjusted current-prompt results are a different population. Metadata control is labeled separately. No per-model best-format aggregation.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Around December 2027. 3 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Muse Spark 1.2 | 79.17 | #64 |
| 2 | GPT-5.6 Sol | 58.33 | #19 |
| 3 | Qwen 3.5 Omni Plus | 16.67 | #98 |
No result here: #4 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=musp-bench-pdf-composer-title-48-questions · How It Works · Data refreshed daily, snapshot 2026-10-22.