MMLU-Redux

An improved version of the MMLU benchmark featuring manually re-annotated questions to identify and correct errors in the original dataset. Provides more reliable evaluation metrics for language models by addressing dataset quality issues found in the original MMLU.

id mmlu-redux · max 1 · 50 models reported

#	Model	Score
1	Qwen3.7 Max Alibaba Cloud / Qwen Team	0.95
2	Qwen3.5-397B-A17B Alibaba Cloud / Qwen Team · open	0.95
3	Qwen3.6 Plus Alibaba Cloud / Qwen Team	0.94
4	Qwen3.7-Plus Alibaba Cloud / Qwen Team	0.94
5	Kimi K2-Thinking-0905 Moonshot AI · open	0.94
6	Qwen3.5-122B-A10B Alibaba Cloud / Qwen Team · open	0.94
7	Qwen3-235B-A22B-Thinking-2507 Alibaba Cloud / Qwen Team · open	0.94
8	Qwen3 VL 235B A22B Thinking Alibaba Cloud / Qwen Team · open	0.94
9	Qwen3.6-27B Alibaba Cloud / Qwen Team · open	0.94
10	DeepSeek-R1-0528 DeepSeek · open	0.93
11	Qwen3.5-35B-A3B Alibaba Cloud / Qwen Team · open	0.93
12	Qwen3.6-35B-A3B Alibaba Cloud / Qwen Team · open	0.93
13	Qwen3.5-27B Alibaba Cloud / Qwen Team · open	0.93
14	Qwen3-235B-A22B-Instruct-2507 Alibaba Cloud / Qwen Team · open	0.93
15	MiMo-V2.5-Pro Xiaomi · open	0.93
16	Qwen3 Max Thinking Alibaba Cloud / Qwen Team	0.93
17	Kimi K2 Instruct Moonshot AI · open	0.93
18	Kimi K2-Instruct-0905 Moonshot AI · open	0.93
19	Qwen3-Next-80B-A3B-Thinking Alibaba Cloud / Qwen Team · open	0.93
20	Qwen3 VL 235B A22B Instruct Alibaba Cloud / Qwen Team · open	0.92
21	Qwen3 VL 32B Thinking Alibaba Cloud / Qwen Team · open	0.92
22	DeepSeek-V3.1 DeepSeek · open	0.92
23	Qwen3.5-9B Alibaba Cloud / Qwen Team · open	0.91
24	Qwen3 VL 30B A3B Thinking Alibaba Cloud / Qwen Team · open	0.91
25	Qwen3-Next-80B-A3B-Instruct Alibaba Cloud / Qwen Team · open	0.91
26	Qwen3 VL 32B Instruct Alibaba Cloud / Qwen Team · open	0.90
27	LongCat-Flash-Thinking Meituan · open	0.89
28	DeepSeek-V3 DeepSeek · open	0.89
29	Qwen3 VL 8B Thinking Alibaba Cloud / Qwen Team · open	0.89
30	Qwen3.5-4B Alibaba Cloud / Qwen Team · open	0.89