GRACE - MuSiQue Categorization: leaderboard

Metric: Four-class error-categorization macro-F1 (%) on MuSiQue grounding traces; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 10 models tracked.

Top models

#ModelScore
1Gemini 3.1 Pro (Preview)69.56
2Gemini 3 Flash56.75
3Gemma 4 31B55.47
4GPT-5.453.53
5Qwen 3.5 27B52.1
6Qwen 3.5 35B A3B51.64
7GPT-5.4 Mini37.16
8Qwen 3 14B35.26
9Qwen 3 8B31.37
10Llama 3.1 8B20.3

Interactive version: theaggregate.ai/benchmark?slug=grace-musique-categorization · How It Works · Data refreshed daily, snapshot 2026-09-29.