GLM 4.6 — Benchmarks

Benchmark scores for GLM 4.6 aggregated from public leaderboards, with how it ranks among open models. See hardware requirements for what you need to run it.

Coding

Benchmark	Score	Open rank	All models
SWE-bench Verified	68.2%	#3 / 13	#50 / 163
SWE-bench Bash Only	55.4%	#5 / 9	#32 / 48
Terminal-Bench	24.5%	#13 / 16	#50 / 57

Math

Benchmark	Score	Open rank	All models
FrontierMath	3.8%	#8 / 12	#76 / 101

Scores aggregated from public benchmark sources (each linked from the benchmark pages). llmrun does not run these benchmarks.