MMLU-Pro leaderboard

An enhanced version of MMLU with 10 answer choices instead of 4, featuring more reasoning-focused questions that better differentiate frontier models.

Models ranked by their published Massive Multitask Language Understanding Professional value

Rank	Model	Provider	10-way multiple choice
1	Qwen3.7 Max	Alibaba	89.6
2	Claude Opus 4.5	Anthropic	89.5
3	Qwen3.6 Plus	Alibaba	88.5
3	Qwen3.7 Plus	Alibaba	88.5
5	Qwen3.5 397B	Alibaba	87.8
6	DeepSeek V4 Pro 0813	DeepSeek	87.5
7	Kimi K2.5	Moonshot AI	87.1
7	Kimi K2.5 (Reasoning)	Moonshot AI	87.1
9	Nemotron 3 Ultra	NVIDIA	86.8
10	Qwen3.5-122B-A10B	Alibaba	86.7
11	Solar Pro 4	Upstage	86.3
12	DeepSeek V4 Flash 0731	DeepSeek	86.2
12	Qwen3.6-27B	Alibaba	86.2
12	Solar Open 2	Upstage	86.2
15	Qwen3.5-27B	Alibaba	86.1
16	GLM-5	Z.AI	85.7
17	Qwen3.5-35B-A3B	Alibaba	85.3
18	Gemma 4 31B	Google	85.2
18	Qwen3.6-35B-A3B	Alibaba	85.2
20	MAI-Thinking-1	Microsoft	85
21	MiMo-V2-Flash	Xiaomi	84.9
22	GLM-4.7	Z.AI	84.3
23	K-EXAONE 2.0	LG AI Research	83.5
24	Qwen3 235B 2507	Alibaba	83
25	Gemma 4 26B A4B	Google	82.6
26	Claude Opus 4.6	Anthropic	82
27	Exaone 4.0 32B	LG AI Research	81.8
28	Nemotron 3.5 Lightning 30B A3B NVFP4	NVIDIA	81.62
29	LongCat-Flash-Lite-Sparse	Meituan	79.24
30	Claude Sonnet 4.6	Anthropic	79.2
31	Granite 4.2 30B	IBM	77.6
32	Nemotron 3 Nano Omni 30B A3B	NVIDIA	77.3
33	Gemma 4 12B	Google	77.2
34	CECeleris-1	Celeris	75.9
34	DeepSeek V3	DeepSeek	75.9
36	ZYZAYA1-8B	Zyphra	74.2
37	Granite 4.2 8B	IBM	74.04
38	OPMiniCPM5-2B	OpenBMB	70.8
39	Gemma 4 E4B	Google	69.4
40	ZYZAYA1-74B-Preview	Zyphra	68.1
41	Granite 4.2 3B	IBM	67.84
42	Gemma 4 E2B	Google	60
43	SPSoofi S 30B-A3B	Soofi Project	51.4
44	OPMiniCPM5-1B	OpenBMB	48.85
45	LFM2.5-230M	LiquidAI	20.25
46	LFM2.5-VL-450M	LiquidAI	19.32

Evidence key: Observed

Rows are ordered by the value MMLU-Pro: A More Robust and Challenging Multi-Task Language Understanding Benchmark published, highest first. The source does not state whether a higher value is the better result, so this page does not either: for a benchmark that measures a rate of failure, read the table from the bottom.