Side-by-side benchmark comparison across knowledge, coding, math, and reasoning.
GLM-4.7 wins overall with a score of 67 vs 21 (46 point difference).GLM-4.7 wins 4 out of 4 categories.
GLM-4.7
83
Mistral 7B v0.3
27.8
GLM-4.7
78
Mistral 7B v0.3
22
GLM-4.7
85
Mistral 7B v0.3
29
GLM-4.7
81
Mistral 7B v0.3
27
GLM-4.7 scores higher overall with 67 vs 21, a difference of 46 points across all benchmarks.
GLM-4.7 leads in knowledge tasks with an average score of 83 vs 27.8.
GLM-4.7 leads in coding with an average score of 78 vs 22.
GLM-4.7 leads in math with an average score of 85 vs 29.
GLM-4.7 leads in reasoning with an average score of 81 vs 27.