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.

Overall rank: #118 of 123 open modelscomposite 17.2/100 across 6 benchmarks in 3 categories · methodology

Coding

BenchmarkScoreOpen rankAll models
Terminal-Bench24.5%#12 / 16#49 / 57
SciCode38.4%#27 / 50#119 / 155
SWE-bench Verified68.2%#7 / 16#50 / 162
SWE-bench Bash Only55.4%#5 / 9#32 / 48

Math

BenchmarkScoreOpen rankAll models
FrontierMath3.8%#8 / 12#76 / 101

Reasoning

BenchmarkScoreOpen rankAll models
CritPt1.1%#26 / 50#109 / 167

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