FMSU-Bench - Rhythm (English): leaderboard

FMSU-Bench (2026) tests fine-grained speech understanding in Chinese and English with multiple-choice questions whose distractors include subtle acoustic confusions and text-only semantic traps, plus a tagged transcription task. This board is accuracy on identifying the rhythmic features in the English subset; percent, higher is better.

Metric: Accuracy (%) of identifying the rhythmic features (RHY), on the FMSU-Bench English subset (500 to 1,000 manually verified items per task, from audiovisual sources), multiple-choice questions whose distractors include fine-grained acoustic confusions and text-only semantic traps; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 10 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Gemini 3.1 Pro (Preview)91.9#99
2Gemini 3 Flash91.6#175
3Gemini 2.5 Flash90.1#463
4Qwen 3 Omni 30B A3B84.4#847
5MiMo-Audio81.5#931
6Kimi-Audio74.5#1735
7Qwen2.5-Omni-7B70.5#1884
8Audio Flamingo 370
9FMSU-Bench Step-Audio 2 (checkpoint unspecified)65.8
10Qwen2-Audio (FMSU-Bench checkpoint unspecified)17.1

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=fmsu-bench-rhythm-english · How It Works · Data refreshed daily, snapshot 2026-11-21.