Side-by-side benchmark comparison across knowledge, coding, math, and reasoning.
Qwen3 235B 2507 (Reasoning) wins overall with a score of 31 vs 24 (7 point difference).Qwen3 235B 2507 (Reasoning) wins 4 out of 4 categories.
DeepSeek V3.1
30.8
Qwen3 235B 2507 (Reasoning)
37.8
DeepSeek V3.1
25
Qwen3 235B 2507 (Reasoning)
32
DeepSeek V3.1
32
Qwen3 235B 2507 (Reasoning)
39
DeepSeek V3.1
30
Qwen3 235B 2507 (Reasoning)
37
Qwen3 235B 2507 (Reasoning) scores higher overall with 31 vs 24, a difference of 7 points across all benchmarks.
Qwen3 235B 2507 (Reasoning) leads in knowledge tasks with an average score of 37.8 vs 30.8.
Qwen3 235B 2507 (Reasoning) leads in coding with an average score of 32 vs 25.
Qwen3 235B 2507 (Reasoning) leads in math with an average score of 39 vs 32.
Qwen3 235B 2507 (Reasoning) leads in reasoning with an average score of 37 vs 30.