FMSU-Bench - Gender (Chinese): 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 gender of the speaker in the Chinese subset; percent, higher is better.

Metric: Accuracy (%) of identifying the gender of the speaker (GEN), on the FMSU-Bench Chinese 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
1Qwen 3 Omni 30B A3B99#847
2Gemini 2.5 Flash97#463
3Qwen2.5-Omni-7B96.9#1884
4FMSU-Bench Step-Audio 2 (checkpoint unspecified)90.2
5Audio Flamingo 389.8
6MiMo-Audio88.4#931
7Gemini 3.1 Pro (Preview)86.3#99
8Gemini 3 Flash85.4#175
9Kimi-Audio79.4#1735
10Qwen2-Audio (FMSU-Bench checkpoint unspecified)46.9

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-gender-chinese · How It Works · Data refreshed daily, snapshot 2026-11-21.