Research · Benchmark profile

MMLU-Pro

A more challenging multiple-choice knowledge and reasoning benchmark derived from MMLU.

Ranking weightedCurrent / rollingActive22 models

Data verified 15 July 2026 · Methodology 1.2.0

Benchmark score on MMLU-Pro

Leaderboard · 22 models

Gemini 3 Pro leads at 89.8%, followed by Qwen3.7-Max (89.6%) and Claude Opus 4.5 (89.5%).

Ranking · 3.5%ResearchCurrent

#1Closed

Gemini 3 Pro

Google

gemini-3-pro

89.8%

Overall #16 · Context 2M tokens

#2Closed

Qwen3.7-Max

Alibaba Cloud

qwen3.7-max

89.6%

Overall #13 · Context 1M tokens

#3Closed

Claude Opus 4.5

Anthropic

claude-opus-4-5

89.5%

Overall #30 · Context 1M tokens

Visual analysis

See the field, not just the table.

Three views of the same sourced leaderboard: model placement, the score curve, and descriptive provider averages. These visuals never alter ranking calculations.

ModelsColours

GoogleAlibaba CloudAnthropicMoonshot AINVIDIAZ.AIXiaomiDeepSeek

Leaderboard bars

20 ranked models · higher is better

View accessible chart data

Accessible data for Leaderboard bars
Model	Rank	Provider	Score
Gemini 3 Pro	#1	Google	89.8%
Qwen3.7-Max	#2	Alibaba Cloud	89.6%
Claude Opus 4.5	#3	Anthropic	89.5%
Qwen3.6 Plus	#4	Alibaba Cloud	88.5%
Qwen3.7-Plus	#5	Alibaba Cloud	88.5%
Qwen3.5 397B	#6	Alibaba Cloud	87.8%
Kimi K2.5	#7	Moonshot AI	87.1%
Nemotron 3 Ultra	#8	NVIDIA	86.8%
Qwen3.5 122B	#9	Alibaba Cloud	86.7%
Qwen3.6 27B	#10	Alibaba Cloud	86.2%
Qwen3.5 27B	#11	Alibaba Cloud	86.1%
GLM-5	#12	Z.AI	85.7%
Qwen3.5 35B	#13	Alibaba Cloud	85.3%
Gemma 4 31B	#14	Google	85.2%
MiMo-V2-Flash	#15	Xiaomi	84.9%
GLM-4.7	#16	Z.AI	84.3%
DeepSeek V4 Flash	#17	DeepSeek	83%
DeepSeek V4 Pro	#18	DeepSeek	82.9%
Gemma 4 26B	#19	Google	82.6%
Claude Opus 4.6	#20	Anthropic	82%

Score curve

Shows how quickly scores separate down the field

View accessible chart data

Accessible data for Score curve
Model	Rank	Provider	Score
Gemini 3 Pro	#1	Google	89.8%
Qwen3.7-Max	#2	Alibaba Cloud	89.6%
Claude Opus 4.5	#3	Anthropic	89.5%
Qwen3.6 Plus	#4	Alibaba Cloud	88.5%
Qwen3.7-Plus	#5	Alibaba Cloud	88.5%
Qwen3.5 397B	#6	Alibaba Cloud	87.8%
Kimi K2.5	#7	Moonshot AI	87.1%
Nemotron 3 Ultra	#8	NVIDIA	86.8%
Qwen3.5 122B	#9	Alibaba Cloud	86.7%
Qwen3.6 27B	#10	Alibaba Cloud	86.2%
Qwen3.5 27B	#11	Alibaba Cloud	86.1%
GLM-5	#12	Z.AI	85.7%
Qwen3.5 35B	#13	Alibaba Cloud	85.3%
Gemma 4 31B	#14	Google	85.2%
MiMo-V2-Flash	#15	Xiaomi	84.9%
GLM-4.7	#16	Z.AI	84.3%
DeepSeek V4 Flash	#17	DeepSeek	83%
DeepSeek V4 Pro	#18	DeepSeek	82.9%
Gemma 4 26B	#19	Google	82.6%
Claude Opus 4.6	#20	Anthropic	82%

Provider profile

Average within the visible cohort · descriptive only

View accessible chart data

Accessible data for Provider profile
Provider	Models	Average score
Alibaba Cloud	8	87.3%
Moonshot AI	1	87.1%
NVIDIA	1	86.8%
Google	3	85.9%
Anthropic	2	85.8%
Z.AI	2	85%
Xiaomi	1	84.9%
DeepSeek	2	83.0%

Benchmark score table (22 models)

One best score per model · higher is better

MMLU-Pro leaderboard ranked by score
Rank	Model	Provider	License	Score
#1	Gemini 3 Pro gemini-3-pro	Google	closed	89.8%
#2	Qwen3.7-Max qwen3.7-max	Alibaba Cloud	—	89.6%
#3	Claude Opus 4.5 claude-opus-4-5	Anthropic	closed	89.5%
#4	Qwen3.6 Plus qwen3-6-plus	Alibaba Cloud	closed	88.5%
#5	Qwen3.7-Plus qwen3.7-plus	Alibaba Cloud	—	88.5%
#6	Qwen3.5 397B qwen3.5-397b-a17b	Alibaba Cloud	open	87.8%
#7	Kimi K2.5 kimi-k2.5	Moonshot AI	—	87.1%
#8	Nemotron 3 Ultra nemotron-3-ultra-550b	NVIDIA	open	86.8%
#9	Qwen3.5 122B qwen3.5-122b-a10b	Alibaba Cloud	open	86.7%
#10	Qwen3.6 27B qwen3.6-27b	Alibaba Cloud	open	86.2%
#11	Qwen3.5 27B qwen3.5-27b	Alibaba Cloud	open	86.1%
#12	GLM-5 glm-5	Z.AI	open	85.7%
#13	Qwen3.5 35B qwen3.5-35b-a3b	Alibaba Cloud	open	85.3%
#14	Gemma 4 31B gemma-4-31b	Google	open	85.2%
#15	MiMo-V2-Flash mimo-v2-flash	Xiaomi	open	84.9%
#16	GLM-4.7 glm-4.7	Z.AI	open	84.3%
#17	DeepSeek V4 Flash deepseek-v4-flash	DeepSeek	open	83%
#18	DeepSeek V4 Pro deepseek-v4-pro	DeepSeek	open	82.9%
#19	Gemma 4 26B gemma-4-26b-a4b	Google	open	82.6%
#20	Claude Opus 4.6 claude-opus-4-6	Anthropic	closed	82%
#21	Claude Sonnet 4.6 claude-sonnet-4-6	Anthropic	closed	79.2%
#22	Gemma 4 12B gemma-4-12b	Google	open	77.2%

The top of this snapshot is led by Gemini 3 Pro at 89.8%; third place is 0.30 points behind. The top-10 spread is 3.6 points.

Compare top models

Gemini 3 Pro vs Qwen3.7-Max Qwen3.7-Max vs Claude Opus 4.5 Claude Opus 4.5 vs Qwen3.6 Plus Qwen3.6 Plus vs Qwen3.7-Plus

About MMLU-Pro

Definition and scoring

Organisation: TIGER-Lab
Category: Research
Version: Not evaluated
Direction: higher is better
Ranking use: Weighted 3.5%
Contamination risk: Unknown

A more challenging multiple-choice knowledge and reasoning benchmark derived from MMLU. Results stay tied to the exact model variant and evaluation system. Multiple systems for the same model use the best published score on this page; overall Lumina scoring uses the median of ranking-eligible rows.

Open benchmark source ↗

FAQ

Quick answers

What does MMLU-Pro measure?

A more challenging multiple-choice knowledge and reasoning benchmark derived from MMLU.

Which model scores highest on MMLU-Pro?

Gemini 3 Pro by Google currently leads with 89.8%.

How many models are evaluated on MMLU-Pro?

22 models in the LuminaBench cohort have a qualifying score on this benchmark.

Does this affect overall Lumina rank?

Yes. This family is ranking-weighted at 3.5% of its capability category. Gemini 3 Pro is overall #16.

SimpleQA

Rolling · 2 results · Weighted 3.5%

PaperBench

Rolling · 2 results · Weighted 3.5%

Finance Agent v2

v2 · 4 results · Not Weighted

BenchLM Knowledge prior

bench-align-v5.1 · 33 results · Weighted 1.6%

AA Long Context Reasoning

Rolling · 96 results · Weighted 1.5%

AA-Omniscience

Rolling · 94 results · Weighted 1.5%

Research · Benchmark profile

MMLU-Pro

A more challenging multiple-choice knowledge and reasoning benchmark derived from MMLU.

Ranking weightedCurrent / rollingActive22 models

Data verified 15 July 2026 · Methodology 1.2.0

Benchmark score on MMLU-Pro

Leaderboard · 22 models

Gemini 3 Pro leads at 89.8%, followed by Qwen3.7-Max (89.6%) and Claude Opus 4.5 (89.5%).

Ranking · 3.5%ResearchCurrent

#1Closed

Gemini 3 Pro

Google

gemini-3-pro

89.8%

Overall #16 · Context 2M tokens

#2Closed

Qwen3.7-Max

Alibaba Cloud

qwen3.7-max

89.6%

Overall #13 · Context 1M tokens

#3Closed

Claude Opus 4.5

Anthropic

claude-opus-4-5

89.5%

Overall #30 · Context 1M tokens

Visual analysis

See the field, not just the table.

Three views of the same sourced leaderboard: model placement, the score curve, and descriptive provider averages. These visuals never alter ranking calculations.

ModelsColours

GoogleAlibaba CloudAnthropicMoonshot AINVIDIAZ.AIXiaomiDeepSeek

Leaderboard bars

20 ranked models · higher is better

View accessible chart data

Accessible data for Leaderboard bars
Model	Rank	Provider	Score
Gemini 3 Pro	#1	Google	89.8%
Qwen3.7-Max	#2	Alibaba Cloud	89.6%
Claude Opus 4.5	#3	Anthropic	89.5%
Qwen3.6 Plus	#4	Alibaba Cloud	88.5%
Qwen3.7-Plus	#5	Alibaba Cloud	88.5%
Qwen3.5 397B	#6	Alibaba Cloud	87.8%
Kimi K2.5	#7	Moonshot AI	87.1%
Nemotron 3 Ultra	#8	NVIDIA	86.8%
Qwen3.5 122B	#9	Alibaba Cloud	86.7%
Qwen3.6 27B	#10	Alibaba Cloud	86.2%
Qwen3.5 27B	#11	Alibaba Cloud	86.1%
GLM-5	#12	Z.AI	85.7%
Qwen3.5 35B	#13	Alibaba Cloud	85.3%
Gemma 4 31B	#14	Google	85.2%
MiMo-V2-Flash	#15	Xiaomi	84.9%
GLM-4.7	#16	Z.AI	84.3%
DeepSeek V4 Flash	#17	DeepSeek	83%
DeepSeek V4 Pro	#18	DeepSeek	82.9%
Gemma 4 26B	#19	Google	82.6%
Claude Opus 4.6	#20	Anthropic	82%

Score curve

Shows how quickly scores separate down the field

View accessible chart data

Accessible data for Score curve
Model	Rank	Provider	Score
Gemini 3 Pro	#1	Google	89.8%
Qwen3.7-Max	#2	Alibaba Cloud	89.6%
Claude Opus 4.5	#3	Anthropic	89.5%
Qwen3.6 Plus	#4	Alibaba Cloud	88.5%
Qwen3.7-Plus	#5	Alibaba Cloud	88.5%
Qwen3.5 397B	#6	Alibaba Cloud	87.8%
Kimi K2.5	#7	Moonshot AI	87.1%
Nemotron 3 Ultra	#8	NVIDIA	86.8%
Qwen3.5 122B	#9	Alibaba Cloud	86.7%
Qwen3.6 27B	#10	Alibaba Cloud	86.2%
Qwen3.5 27B	#11	Alibaba Cloud	86.1%
GLM-5	#12	Z.AI	85.7%
Qwen3.5 35B	#13	Alibaba Cloud	85.3%
Gemma 4 31B	#14	Google	85.2%
MiMo-V2-Flash	#15	Xiaomi	84.9%
GLM-4.7	#16	Z.AI	84.3%
DeepSeek V4 Flash	#17	DeepSeek	83%
DeepSeek V4 Pro	#18	DeepSeek	82.9%
Gemma 4 26B	#19	Google	82.6%
Claude Opus 4.6	#20	Anthropic	82%

Provider profile

Average within the visible cohort · descriptive only

View accessible chart data

Accessible data for Provider profile
Provider	Models	Average score
Alibaba Cloud	8	87.3%
Moonshot AI	1	87.1%
NVIDIA	1	86.8%
Google	3	85.9%
Anthropic	2	85.8%
Z.AI	2	85%
Xiaomi	1	84.9%
DeepSeek	2	83.0%

Benchmark score table (22 models)

One best score per model · higher is better

MMLU-Pro leaderboard ranked by score
Rank	Model	Provider	License	Score
#1	Gemini 3 Pro gemini-3-pro	Google	closed	89.8%
#2	Qwen3.7-Max qwen3.7-max	Alibaba Cloud	—	89.6%
#3	Claude Opus 4.5 claude-opus-4-5	Anthropic	closed	89.5%
#4	Qwen3.6 Plus qwen3-6-plus	Alibaba Cloud	closed	88.5%
#5	Qwen3.7-Plus qwen3.7-plus	Alibaba Cloud	—	88.5%
#6	Qwen3.5 397B qwen3.5-397b-a17b	Alibaba Cloud	open	87.8%
#7	Kimi K2.5 kimi-k2.5	Moonshot AI	—	87.1%
#8	Nemotron 3 Ultra nemotron-3-ultra-550b	NVIDIA	open	86.8%
#9	Qwen3.5 122B qwen3.5-122b-a10b	Alibaba Cloud	open	86.7%
#10	Qwen3.6 27B qwen3.6-27b	Alibaba Cloud	open	86.2%
#11	Qwen3.5 27B qwen3.5-27b	Alibaba Cloud	open	86.1%
#12	GLM-5 glm-5	Z.AI	open	85.7%
#13	Qwen3.5 35B qwen3.5-35b-a3b	Alibaba Cloud	open	85.3%
#14	Gemma 4 31B gemma-4-31b	Google	open	85.2%
#15	MiMo-V2-Flash mimo-v2-flash	Xiaomi	open	84.9%
#16	GLM-4.7 glm-4.7	Z.AI	open	84.3%
#17	DeepSeek V4 Flash deepseek-v4-flash	DeepSeek	open	83%
#18	DeepSeek V4 Pro deepseek-v4-pro	DeepSeek	open	82.9%
#19	Gemma 4 26B gemma-4-26b-a4b	Google	open	82.6%
#20	Claude Opus 4.6 claude-opus-4-6	Anthropic	closed	82%
#21	Claude Sonnet 4.6 claude-sonnet-4-6	Anthropic	closed	79.2%
#22	Gemma 4 12B gemma-4-12b	Google	open	77.2%

The top of this snapshot is led by Gemini 3 Pro at 89.8%; third place is 0.30 points behind. The top-10 spread is 3.6 points.

Compare top models

Gemini 3 Pro vs Qwen3.7-Max Qwen3.7-Max vs Claude Opus 4.5 Claude Opus 4.5 vs Qwen3.6 Plus Qwen3.6 Plus vs Qwen3.7-Plus

About MMLU-Pro

Definition and scoring

Organisation: TIGER-Lab
Category: Research
Version: Not evaluated
Direction: higher is better
Ranking use: Weighted 3.5%
Contamination risk: Unknown

Open benchmark source ↗

FAQ

Quick answers

What does MMLU-Pro measure?

A more challenging multiple-choice knowledge and reasoning benchmark derived from MMLU.

Which model scores highest on MMLU-Pro?

Gemini 3 Pro by Google currently leads with 89.8%.

How many models are evaluated on MMLU-Pro?

22 models in the LuminaBench cohort have a qualifying score on this benchmark.

Does this affect overall Lumina rank?

Yes. This family is ranking-weighted at 3.5% of its capability category. Gemini 3 Pro is overall #16.

SimpleQA

Rolling · 2 results · Weighted 3.5%

PaperBench

Rolling · 2 results · Weighted 3.5%

Finance Agent v2

v2 · 4 results · Not Weighted

BenchLM Knowledge prior

bench-align-v5.1 · 33 results · Weighted 1.6%

AA Long Context Reasoning

Rolling · 96 results · Weighted 1.5%

AA-Omniscience

Rolling · 94 results · Weighted 1.5%

Leaderboard · 22 models

Gemini 3 Pro

Qwen3.7-Max

Claude Opus 4.5

See the field, not just the table.

Benchmark score table (22 models)

Compare top models

Definition and scoring

Quick answers

What does MMLU-Pro measure?

Which model scores highest on MMLU-Pro?

How many models are evaluated on MMLU-Pro?

Does this affect overall Lumina rank?

More Research benchmarks

Leaderboard · 22 models

Gemini 3 Pro

Qwen3.7-Max

Claude Opus 4.5

See the field, not just the table.

Benchmark score table (22 models)

Compare top models

Definition and scoring

Quick answers

What does MMLU-Pro measure?

Which model scores highest on MMLU-Pro?

How many models are evaluated on MMLU-Pro?

Does this affect overall Lumina rank?

More Research benchmarks