Reasoning Trajectory Evaluation on MentraBench 1.0 (test) benchmark leaderboard · SOTA2 Research

Reasoning Trajectory Evaluation on MentraBench 1.0 (test)

100Conciseness ScoreGPT-o4-mini

Evaluation Results

Method	Links
GPT-o4-mini2025.12		100	0.4135	100	0.9519	100	0.8731
LLaMA 42025.12		100	0.2596	100	0.8942	94.23	0.8192
Deepseek-R12025.12		95.19	0.9808	100	0.9808	100	0.9827
Psyche-R12025.12		94.23	0.9038	100	0.9712	90.38	0.9442
Mindora_CHORD2025.12		94.23	0.9519	100	0.9904	98.08	0.9731
Qwen plus2025.12		65.38	1	100	0.9808	100	0.927
Qwen3-8B2025.12		57.69	0.9712	100	0.9808	99.04	0.9039