Terminal-Bench 2.0 ↗24.53%
GLM-4.6 · Terminal Bench · Accuracy mean · % ↗
1 more reported result
24.5% · GLM-4.6 · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗
Z.ai / MODEL EVIDENCE
GLM-4.6
16 benchmarks · 17 observations · 17 sources · released 2025-09-30
Scores at a glance
16 benchmarks · large dot: GLM-4.6; small dots: closest models where reported
CritPt ↗1.14%
GLM-4.6 · CritPt · Accuracy · % ↗
SciCode ↗38.43%
GLM-4.6 · SciCode · Score · % ↗
Text Arena (Coding) ↗1340.6 Elo
GLM-4.6 · WebDev Arena · Elo ↗
LegalBench ↗79.61 %
GLM 4.6 · LegalBench · Accuracy ↗
CorpFin v2 ↗56.84 %
GLM 4.6 · CorpFin · Accuracy ↗
MortgageTax & TaxEval v2 ↗66.23 %
GLM 4.6 · TaxEval · Accuracy ↗
MedQA (USMLE) ↗92.22 %
GLM 4.6 · MedQA · Accuracy ↗
AIME 2024–2025 · Vals ↗92.71 %
GLM 4.6 · AIME · Accuracy ↗
MGSM (Multilingual GSM) ↗89.75 %
GLM 4.6 · MGSM · Accuracy ↗
GPQA Diamond ↗74.5 %
GLM 4.6 · GPQA Diamond · Accuracy ↗
MMLU-Pro ↗82.2 %
GLM 4.6 · MMLU Pro · Accuracy ↗
Vibe Code Bench ↗3.09 %
GLM 4.6 (OpenHands) · Vibe Code Bench · original task set ↗
LiveCodeBench ↗81.04 %
GLM 4.6 · LiveCodeBench · Accuracy ↗
Poker Agent Benchmark ↗945.76 TrueSkill
GLM 4.6 · Poker Agent · TrueSkill Rating ↗
ALE-Bench ↗340.82 Performance points
glm-4.6 · Average contest performance ↗
Closest models by score
Similar positions on shared benchmark leaderboards. Smaller gaps are closer.
MiniMax M2.5 →MiniMax
8 of 11 shared benchmark comparisons
GLM 4.7 →Z.ai
9 of 11 shared benchmark comparisons
GPT-5.4 Nano →OpenAI
7 of 11 shared benchmark comparisons
How similarity is measured
Mean absolute gap in percentile positions on shared benchmark settings; lower is closer. At least 60% overlap is required, with a minimum of three when available. Independent evaluations are used when there are at least three; otherwise published report comparisons are used. Repeated model observations within one setting are averaged before ranking. Writing-style traits and composite indices are excluded. This describes similarity, not an overall quality ranking.
Sources
Every plotted observation links to its evaluation. Reported configurations, benchmark versions and task subsets can differ; similarity describes shared scores, not training lift.
- Arena leaderboard — WebDev evaluation ↗
- SciCode Benchmark | Artificial Analysis ↗
- Ale Bench — methodology and published results ↗
- CritPt v1 ↗
- Terminal-Bench 2.0 ↗
- Terminal-Bench 2.0 leaderboard ↗
- AIME — evaluation methodology and results ↗
- CorpFin — evaluation methodology and results ↗
- GPQA Diamond — evaluation methodology and results ↗
- LiveCodeBench — evaluation methodology and results ↗
- LegalBench — evaluation methodology and results ↗
- MedQA — evaluation methodology and results ↗
- MGSM — evaluation methodology and results ↗
- MMLU Pro — evaluation methodology and results ↗
- Poker Agent — evaluation methodology and results ↗
- TaxEval — evaluation methodology and results ↗
- Vibe Code Bench — evaluation methodology and results ↗
Names used in source tables
GLM 4.6 · GLM 4.6 (OpenHands) · GLM-4.6 · glm-4.6