Successor to GLM 5.1. AA Intelligence Index 50.7, GPQA 89.5, Terminal-Bench 50.8 — open-weight reasoning gains over 5.1.
SI v3 uses weighted percentages from one Artificial Analysis configuration. Ranking requires all five benchmarks. Partial scores average only available weights. Scores are not comparable to SI v2 or a probability of reaching singularity.
Compared with models released within six months of it.
| Benchmark | Score | Rank |
|---|---|---|
TerminalArtificial Analysis Agentic terminal coding tasks requiring multi-step execution | 50.8% | #19 / 58 |
MMLU-Provals.ai Harder 10-option successor to MMLU; more reasoning-focused | 86.7% | #37 / 61 |
HLEArtificial Analysis Challenging multidisciplinary questions evaluated by Artificial Analysis | 41.1% | #39 / 102 |
GPQAArtificial Analysis PhD-level science questions even experts struggle with | 89.5% | #45 / 96 |
LiveCodeBenchvals.ai Contamination-free competitive programming (filtered by cutoff date) | 69.5% | #55 / 62 |
| Benchmark / source | Score |
|---|---|
HLE Artificial AnalysisGLM-5.2 (Max) | 41.1% |
Terminal-Bench 2.1 Artificial AnalysisGLM-5.2 (Max) | 77.9% |
Terminal-Bench 4.0 Artificial AnalysisGLM-5.2 (Max) | 1% |
SciCode Artificial AnalysisGLM-5.2 (Max) | 51.2% |
AA-LCR Artificial AnalysisGLM-5.2 (Max) | 78.3% |
CritPt Artificial AnalysisGLM-5.2 (Max) | 20.9% |
ITBench SRE Artificial AnalysisGLM-5.2 (Max) | 42.7% |
IFBench Artificial AnalysisGLM-5.2 (Max) | 73.3% |