GLM-5.1: benchmark results
Zhipu GLM-5.1 model for reasoning, coding, and agentic tasks. Provider: Zhipu. Released 2026-03-27. Access: Open.
Aggregate Rating 1763 ± 3, rank #97 of 1656 rated models, from 1261 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AGC-Bench - cue_word_story | 1.25 | Dataset z-score | 100 |
| AGC-Bench - simile_generation | 1.9 | Dataset z-score | 100 |
| AGORA - Finance | 56.41 | Accuracy (%) on the 39 finance and economics questions (5,96 | 100 |
| AhaBench - Vending (Incident) | 2884 | Mean profit in US dollars over five seeds (incident setting | 100 |
| Are Agents Ready to Teach? A Multi-Stage Bench | 63.8 | Eq. pass (self-reported) | 100 |
| BoxPwnr CTF Bench | 55.47 | Mean published platform score (%); observed platforms only; | 100 |
| CHI-Bench (DeepAgents) | 11.11 | Pass@1 (%) | 100 |
| CUE-Bench - Fine-grained Emotion | 30.5 | Accuracy (%; fine-grained emotion classification; over CUE-B | 100 |
| Claw-SWE-Bench (Hermes Agent) | 71.1 | Pass@1 (%) on the 350 GitHub issue-resolution instances of C | 100 |
| Claw-SWE-Bench (nanobot) | 60.9 | Pass@1 (%) on the 350 GitHub issue-resolution instances of C | 100 |
| ConstraintRot | 0 | Violation rate (%; share of 27 episodes, 9 tasks x 3 repetit | 100 |
| EIBench (Emotion Management) - Repair | 19.3 | Anchor-based outcome score x 100 (-100 to +100): the simulat | 100 |
Interactive version: theaggregate.ai/model?slug=glm-5-1 · How It Works · Data refreshed daily, snapshot 2026-10-20.