MMLU-ProX

MMLU-ProX - multilingual extended MMLU benchmark, average across en/de/fr/es/it/ja/zh/hi/pt/ko.

Category

Multilingual

Max Score

100.0

Score Type

percent

Active

Yes

Model Scores

Scores for MMLU-ProX

Model Score Date Verified
MiniMax-M2.7
48.39%
04.06.2026 Unverified
GLM-5.1
96.13%
04.06.2026 Unverified
Kimi K2.6
90.97%
04.06.2026 Unverified
Qwen3.5-397B-A17B
100.00%
04.06.2026 Unverified
DeepSeek-V4-Pro
94.84%
04.06.2026 Unverified
DeepSeek-V4-Flash
86.45%
04.06.2026 Unverified
Qwen3.5-4B
3.87%
01.02.2026 Verified
Qwen3.5-122B-A10B
72.90%
01.02.2026 Verified
Qwen3.5-27B
72.90%
01.02.2026 Unverified
Qwen3.5-35B-A3B
65.16%
01.02.2026 Unverified
Qwen3-235B-A22B
45.16%
01.02.2026 Unverified
GPT-5-mini 2025-08-07
49.03%
01.02.2026 Unverified
Gemini-2.5-Flash-Thinking
60.00%
31.07.2025 Verified
Qwen3-30B-A3B
14.19%
29.04.2025 Verified
Qwen3-30B-A3B-Thinking-2507
35.48%
31.07.2025 Verified
Qwen3-Next-80B-A3B-Thinking
50.32%
10.09.2025 Verified
Qwen3-VL-30B-A3B
70.90
04.10.2025 Verified
NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16
78.06%
04.06.2026 Verified