Side-by-side benchmark comparison across knowledge, coding, math, and reasoning.
Llama 4 Scout wins overall with a score of 38 vs 22 (16 point difference).Llama 4 Scout wins 4 out of 4 categories.
GPT-OSS 20B
28.8
Llama 4 Scout
44.8
GPT-OSS 20B
23
Llama 4 Scout
39
GPT-OSS 20B
30
Llama 4 Scout
46
GPT-OSS 20B
28
Llama 4 Scout
44
Llama 4 Scout scores higher overall with 38 vs 22, a difference of 16 points across all benchmarks.
Llama 4 Scout leads in knowledge tasks with an average score of 44.8 vs 28.8.
Llama 4 Scout leads in coding with an average score of 39 vs 23.
Llama 4 Scout leads in math with an average score of 46 vs 30.
Llama 4 Scout leads in reasoning with an average score of 44 vs 28.