Claude Haiku 4.5 vs DeepSeek V3.2 (Thinking)

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

Quick Verdict

DeepSeek V3.2 (Thinking) wins overall with a score of 69 vs 57 (12 point difference).DeepSeek V3.2 (Thinking) wins 4 out of 4 categories.

Knowledge

DeepSeek V3.2 (Thinking)

Claude Haiku 4.5

65.8

DeepSeek V3.2 (Thinking)

84

68
MMLU
87
67
GPQA
85
65
SuperGPQA
83
63
OpenBookQA
81

Coding

DeepSeek V3.2 (Thinking)

Claude Haiku 4.5

60

DeepSeek V3.2 (Thinking)

79

60
HumanEval
79

Mathematics

DeepSeek V3.2 (Thinking)

Claude Haiku 4.5

67

DeepSeek V3.2 (Thinking)

86

68
AIME 2023
87
70
AIME 2024
89
69
AIME 2025
88
64
HMMT Feb 2023
83
66
HMMT Feb 2024
85
65
HMMT Feb 2025
84
67
BRUMO 2025
86

Reasoning

DeepSeek V3.2 (Thinking)

Claude Haiku 4.5

64

DeepSeek V3.2 (Thinking)

82

65
SimpleQA
83
63
MuSR
81

Frequently Asked Questions

Which is better, Claude Haiku 4.5 or DeepSeek V3.2 (Thinking)?

DeepSeek V3.2 (Thinking) scores higher overall with 69 vs 57, a difference of 12 points across all benchmarks.

Which is better for knowledge tasks, Claude Haiku 4.5 or DeepSeek V3.2 (Thinking)?

DeepSeek V3.2 (Thinking) leads in knowledge tasks with an average score of 84 vs 65.8.

Which is better for coding, Claude Haiku 4.5 or DeepSeek V3.2 (Thinking)?

DeepSeek V3.2 (Thinking) leads in coding with an average score of 79 vs 60.

Which is better for math, Claude Haiku 4.5 or DeepSeek V3.2 (Thinking)?

DeepSeek V3.2 (Thinking) leads in math with an average score of 86 vs 67.

Which is better for reasoning, Claude Haiku 4.5 or DeepSeek V3.2 (Thinking)?

DeepSeek V3.2 (Thinking) leads in reasoning with an average score of 82 vs 64.