Which model leads MMLU-ProX?

Q: Which model leads MMLU-ProX?

Qwen3.7 Max tops the board at 87

MMLU-ProX Leaderboard — Model Scores

Pro Mode — raw scores, one benchmark under the lensofficial card harness

The field · 36 measured

34.687

36 models measured, most on the official card harness. Qwen3.7 Max tops the board at 87.

36 measured·Lifecycle Updated Oct 8, 2026Score basis

#◆Vendor$/M in$/M out

Leader — ignites whiteField — fill cools with scoreIndependently verified only9 of 36Bars show 34.6–87 · ◆ solid = first-party or better · outlined = arm's-length

01Qwen3.7 MaxAlibaba87$2.5/M$7.5/M$0.057

02Qwen3.7 Plus PreviewAlibaba85.4$0.40/M$1.6/M$0.012

03Qwen3.6 PlusAlibaba84.7$0.50/M$3.0/M$0.021

04Qwen3.5 397B A17BAlibaba84.7$0.60/M$3.6/M$0.025

05Nemotron 3 Ultra 550B A55BNVIDIA83$0.60/M$2.5/M$0.019

06Qwen3.5 122B A10BAlibaba82.2$0.40/M$3.2/M$0.022

07Qwen3.5 27BAlibaba82.2$0.30/M$2.4/M$0.016

08Qwen3.5 35B A3BAlibaba81$0.25/M$2.0/M$0.014

09Qwen3 VL 235B A22B ReasoningAlibaba80.6$0.40/M$4.0/M$0.027

10Qwen3 235B A22B Instruct 2507Alibaba79.4$0.23/M$0.92/M$0.007

11Nemotron 3 Super 120B A12BNVIDIA79.36$0.30/M$0.90/M$0.008

12Qwen3 Next 80B A3BAlibaba78.7$0.15/M$1.2/M$0.009

13Qwen3 VL 235B A22B InstructAlibaba77.8$0.40/M$1.6/M$0.013

14Qwen3 VL 32B ReasoningAlibaba77.2$0.16/M$0.64/M$0.005

15Qwen3 Next 80B A3B InstructAlibaba76.7$0.15/M$1.2/M$0.009

16Qwen3.5 9BAlibaba76.3$0.14/M$0.20/M$0.002

17Qwen3 VL 30B A3B ReasoningAlibaba76.1$0.20/M$2.4/M$0.017

18DeepSeek-R1DeepSeek75.53P$2.0/M$4.0/M$0.040

19Qwen3 VL 32B InstructAlibaba73.4$0.16/M$0.64/M$0.005

20GPT-4.1OpenAI72.73P$2.0/M$8.0/M$0.069

21Qwen3.5 4BAlibaba71.5$0.03/M$0.15/M$0.001

22Qwen3 VL 30B A3B InstructAlibaba70.9$0.20/M$0.80/M$0.007

23Qwen3 VL Thinking (8B)Alibaba70.7$0.18/M$2.1/M$0.016

24DeepSeek-V3DeepSeek70.53P$0.24/M$0.90/M$0.008

25O4 MiniOpenAI69.33P$1.1/M$4.4/M$0.040

26Qwen3 235B A22BAlibaba66.73P$0.70/M$2.8/M$0.026

27Qwen3 VL 8B InstructAlibaba65.4$0.18/M$0.70/M$0.007

28Qwen3 VL 4B (Reasoning)Alibaba65———

29GPT-4oOpenAI61.13P$5.0/M$15.0/M$0.164

30Qwen3 32BAlibaba59.93P$0.16/M$0.64/M$0.007

31NVIDIA Nemotron 3 Nano 30B A3BNVIDIA59.5$0.05/M$0.20/M$0.002

32Qwen3 VL 4B InstructAlibaba59.4———

33Gemma 3 27BGoogle56.63P$0.08/M$0.16/M$0.002

34Qwen3.5 2BAlibaba52.3———

35Phi 4Microsoft49.93P$0.13/M$0.50/M$0.006

36Qwen3.5 0.8BAlibaba34.6———

◆ solid = independently verified · aggregator attested · vendor attributed — outlined = cross-reference · source-attributed. Dimmed ◆ = confidence inferred. “+N alts” = alternate disclosures in the drawer with why-demoted. Prices per million tokens; $/point = input price ÷ score. Every row opens its model dossier.

Verification: 36 models scored · 9 independently verified · 27 vendor-reported · 0 with source disagreement. How these tiers are assigned