Side-by-side benchmark comparison across knowledge, coding, math, and reasoning.
Qwen3.5 397B wins overall with a score of 65 vs 25 (40 point difference).Qwen3.5 397B wins 4 out of 4 categories.
DeepSeek V3.1 (Reasoning)
31.8
Qwen3.5 397B
80.8
DeepSeek V3.1 (Reasoning)
26
Qwen3.5 397B
75
DeepSeek V3.1 (Reasoning)
33
Qwen3.5 397B
82
DeepSeek V3.1 (Reasoning)
31
Qwen3.5 397B
79
Qwen3.5 397B scores higher overall with 65 vs 25, a difference of 40 points across all benchmarks.
Qwen3.5 397B leads in knowledge tasks with an average score of 80.8 vs 31.8.
Qwen3.5 397B leads in coding with an average score of 75 vs 26.
Qwen3.5 397B leads in math with an average score of 82 vs 33.
Qwen3.5 397B leads in reasoning with an average score of 79 vs 31.