Side-by-side benchmark comparison across knowledge, coding, math, and reasoning.
Qwen3.5 397B wins overall with a score of 65 vs 50 (15 point difference).Qwen3.5 397B wins 4 out of 4 categories.
GPT-4 Turbo
58.5
Qwen3.5 397B
80.8
GPT-4 Turbo
52
Qwen3.5 397B
75
GPT-4 Turbo
59
Qwen3.5 397B
82
GPT-4 Turbo
57
Qwen3.5 397B
79
Qwen3.5 397B scores higher overall with 65 vs 50, a difference of 15 points across all benchmarks.
Qwen3.5 397B leads in knowledge tasks with an average score of 80.8 vs 58.5.
Qwen3.5 397B leads in coding with an average score of 75 vs 52.
Qwen3.5 397B leads in math with an average score of 82 vs 59.
Qwen3.5 397B leads in reasoning with an average score of 79 vs 57.