MMLU-Pro
MMLU-Pro - extended MMLU covering 14 categories with complex reasoning questions.
Category
Knowledge
Max Score
100.0
Score Type
percent
Active
Yes
Model Scores
Scores for MMLU-Pro
| Model | Score | Date | Verified |
|---|---|---|---|
| Qwen3.6-35B-A3B |
81.29%
|
16.04.2026 | Verified |
| Gemma4-31B-it |
81.29%
|
16.04.2026 | Unverified |
| Qwen3.5-35B-A3B |
81.61%
|
16.04.2026 | Unverified |
| Gemma4-26B-A4B |
72.90%
|
16.04.2026 | Unverified |
| Qwen3.6-27B |
84.52%
|
22.04.2026 | Verified |
| Claude 4.5 Opus |
95.16%
|
22.04.2026 | Unverified |
| Qwen3.5-27B |
84.19%
|
16.04.2026 | Unverified |
| NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 |
86.45%
|
04.06.2026 | Verified |
| MiniMax-M2.7 |
70.65%
|
04.06.2026 | Unverified |
| GLM-5.1 |
83.55%
|
04.06.2026 | Unverified |
| Kimi K2.6 |
90.65%
|
04.06.2026 | Unverified |
| DeepSeek-V4-Pro |
88.71%
|
04.06.2026 | Unverified |
| Qwen3.5-4B |
61.61%
|
01.02.2026 | Verified |
| Qwen3.5-122B-A10B |
86.13%
|
01.02.2026 | Verified |
| Qwen3-235B-A22B |
78.71%
|
01.02.2026 | Unverified |
| GPT-5-mini 2025-08-07 |
76.45%
|
01.02.2026 | Unverified |
| Gemma4-12B |
55.48%
|
02.07.2026 | Verified |
| Gemma4-E4B |
30.32%
|
02.07.2026 | Verified |
| Gemma4-E2B |
60.00
|
02.07.2026 | Verified |
| Gemma3-27B |
24.52%
|
02.07.2026 | Unverified |
| Gemini-2.5-Flash-Thinking |
70.65%
|
31.07.2025 | Verified |
| Qwen3-30B-A3B |
59.68%
|
29.04.2025 | Verified |
| Qwen3-30B-A3B-Thinking-2507 |
67.42%
|
31.07.2025 | Verified |
| DeepSeek-V4-Flash |
84.52%
|
26.06.2026 | Verified |
| DeepSeek-V3.2 |
17.74%
|
26.06.2026 | Verified |
| DeepSeek-V4-Pro |
43.55%
|
26.06.2026 | Verified |
| Opus-4.6 |
93.87%
|
26.06.2026 | Verified |
| GPT-5.4 |
88.71%
|
26.06.2026 | Verified |
| Gemini 3.1 Pro |
100.00%
|
26.06.2026 | Verified |
| MiniCPM5-2B |
34.84%
|
— | Verified |
| LFM2.5-2.6B |
16.77%
|
— | Verified |
| Qwen3.5-2B |
13.87%
|
— | Verified |
| granite-4.2-3B |
18.71%
|
— | Verified |
| Nemotron-3-Nano-4B |
18.39%
|
— | Verified |
| LFM2.5-8B-A1B |
10.00%
|
— | Verified |
| Qwen3-Next-80B-A3B-Thinking |
73.23%
|
10.09.2025 | Verified |
| Qwen3-VL-30B-A3B |
57.42%
|
04.10.2025 | Verified |
| Qwen3.5-397B-A17B |
89.68%
|
22.04.2026 | Unverified |