DeepSeekMath V2 vs Qwen3.5 397B (Reasoning)

Side-by-side benchmark comparison across knowledge, coding, math, and reasoning.

Quick Verdict

Qwen3.5 397B (Reasoning) wins overall with a score of 70 vs 64 (6 point difference).Qwen3.5 397B (Reasoning) wins 4 out of 4 categories.

Knowledge

Qwen3.5 397B (Reasoning)

DeepSeekMath V2

77.8

Qwen3.5 397B (Reasoning)

88

80
MMLU
91
79
GPQA
89
77
SuperGPQA
87
75
OpenBookQA
85

Coding

Qwen3.5 397B (Reasoning)

DeepSeekMath V2

72

Qwen3.5 397B (Reasoning)

83

72
HumanEval
83

Mathematics

Qwen3.5 397B (Reasoning)

DeepSeekMath V2

79

Qwen3.5 397B (Reasoning)

92

80
AIME 2023
93
82
AIME 2024
95
81
AIME 2025
94
76
HMMT Feb 2023
89
78
HMMT Feb 2024
91
77
HMMT Feb 2025
90
79
BRUMO 2025
92

Reasoning

Qwen3.5 397B (Reasoning)

DeepSeekMath V2

76

Qwen3.5 397B (Reasoning)

86

77
SimpleQA
87
75
MuSR
85

Frequently Asked Questions

Which is better, DeepSeekMath V2 or Qwen3.5 397B (Reasoning)?

Qwen3.5 397B (Reasoning) scores higher overall with 70 vs 64, a difference of 6 points across all benchmarks.

Which is better for knowledge tasks, DeepSeekMath V2 or Qwen3.5 397B (Reasoning)?

Qwen3.5 397B (Reasoning) leads in knowledge tasks with an average score of 88 vs 77.8.

Which is better for coding, DeepSeekMath V2 or Qwen3.5 397B (Reasoning)?

Qwen3.5 397B (Reasoning) leads in coding with an average score of 83 vs 72.

Which is better for math, DeepSeekMath V2 or Qwen3.5 397B (Reasoning)?

Qwen3.5 397B (Reasoning) leads in math with an average score of 92 vs 79.

Which is better for reasoning, DeepSeekMath V2 or Qwen3.5 397B (Reasoning)?

Qwen3.5 397B (Reasoning) leads in reasoning with an average score of 86 vs 76.