Side-by-side benchmark comparison across knowledge, coding, math, and reasoning.
DeepSeek Coder 2.0 wins overall with a score of 64 vs 48 (16 point difference).DeepSeek Coder 2.0 wins 4 out of 4 categories.
DeepSeek Coder 2.0
77.8
Llama 3 70B
56.5
DeepSeek Coder 2.0
82
Llama 3 70B
50
DeepSeek Coder 2.0
80
Llama 3 70B
57
DeepSeek Coder 2.0
77
Llama 3 70B
55
DeepSeek Coder 2.0 scores higher overall with 64 vs 48, a difference of 16 points across all benchmarks.
DeepSeek Coder 2.0 leads in knowledge tasks with an average score of 77.8 vs 56.5.
DeepSeek Coder 2.0 leads in coding with an average score of 82 vs 50.
DeepSeek Coder 2.0 leads in math with an average score of 80 vs 57.
DeepSeek Coder 2.0 leads in reasoning with an average score of 77 vs 55.