Qwen3-Max-Thinking Benchmark Details

Qwen3-Max-Thinking currently shows benchmark results led by C-Eval (1 / 48, score 93.70), HLE (54 / 565, score 49.80), LiveCodeBench (26 / 251, score 85.90).

Benchmark Results

Qwen3-Max-Thinking

Benchmark Results

General Knowledge

6 evaluations

Benchmark / mode

Score

Rank/total

C-Eval

Thinking Mode

93.70

1 / 48

MMLU-Pro

Thinking Mode

85.70

24 / 176

HLE

Thinking Mode

30.20

203 / 565

HLE

Thinking Mode

226 / 565

HLE

Thinking ModeTools

49.80

54 / 565

CritPt

Thinking Mode

1.70

132 / 201

Other

1 evaluations

Benchmark / mode

Score

Rank/total

GPQA Diamond

Thinking Mode

87.40

127 / 463

Coding and Software Engineer

2 evaluations

Benchmark / mode

Score

Rank/total

LiveCodeBench

Thinking Mode

85.90

26 / 251

SWE-bench Verified

Thinking Mode

75.30

38 / 116

Math and Reasoning

2 evaluations

Benchmark / mode

Score

Rank/total

IMO-AnswerBench

Thinking Mode

83.90

13 / 24

AIME2025

Thinking Mode

82.30

91 / 216

Agent Level Benchmark

3 evaluations

Benchmark / mode

Score

Rank/total

τ²-Bench - Telecom

Thinking ModeTools

83.60

101 / 264

τ²-Bench

Thinking ModeTools

82.10

11 / 44

Terminal Bench Hard

Thinking ModeTools

24.20

134 / 244

Instruction Following

2 evaluations

Benchmark / mode

Score

Rank/total

IF Bench

Thinking Mode

70.70

65 / 282

IF Bench

Thinking ModeTools

70.90

64 / 282

Claw-style Agent Evaluation

1 evaluations

Benchmark / mode

Score

Rank/total

Pinch Bench

Thinking ModeTools

80.30

24 / 38

Compare with other models