GLM-4.5 (Non-reasoning): benchmark results
Provider: Zhipu. Released 2025-07-28. Access: Open.
Aggregate Rating 1615 ± 27, rank #668 of 2170 rated models, from 33 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| GMP - Co-occurring Violations | 58 | Macro-F1 (x100) over the 12 violation categories of GMP Task | 100 |
| UGI Leaderboard | 58.64 | UGI Score | 98.3 |
| ReLE - Agents - BFCL-V3 | 70.8 | Accuracy (%) | 89.8 |
| UGI - Writing | 54.18 | Writing Score | 89 |
| GMP - Co-occurring Violations (Micro-F1) | 83 | Micro-F1 (x100) over the 12 violation categories of GMP Task | 88.9 |
| ReLE - Agents and Tool Use | 67.9 | Accuracy (%) | 86.5 |
| UGI - Natural Intelligence | 38.72 | NatInt Score | 84.5 |
| AgentNoiseBench (Search, Tool Noise) | 36 | Accuracy on 500 sampled 2WikiMultiHopQA and HotpotQA questio | 78.3 |
| ReLE - Education - Primary School Subjects | 60.7 | Accuracy (%) | 58.2 |
| AgentNoiseBench (tau2-bench, Tool Noise) | 42 | Success on a stratified 25 percent sample of tau2-bench task | 56.5 |
| ReLE - Reasoning - BBH | 79.5 | Accuracy (%) | 51.7 |
| AgentNoiseBench (tau2-bench, User Noise) | 55 | Success on a stratified 25 percent sample of tau2-bench task | 47.8 |
Interactive version: theaggregate.ai/model?slug=glm-4-5-non-reasoning · How It Works · Data refreshed daily, snapshot 2026-11-05.