GLM-4.7: benchmark results
Zhipu AI's GLM-4.7, a 355B sparse MoE (32B active) flagship reasoning model (December 2025). Provider: Zhipu. Released 2025-12-22. Access: Open.
Unified ELO 1631 ± 1, rank #178 of 1392 rated models, from 282 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AGC-Bench - historical_analogy | 1.94 | Dataset z-score | 99.4 |
| FormationEval | 98.6 | Accuracy (%) | 98.6 |
| AGC-Bench - c3_crosstalk | 1.42 | Dataset z-score | 97.6 |
| CLEM Hot Air Balloon | 90.89 | Game Clemscore (%) | 95.8 |
| AGC-Bench - arn | 0.86 | Dataset z-score | 95.7 |
| AGC-Bench - schnovel | 1.56 | Dataset z-score | 95.7 |
| OpenCompass Research - LiveCodeBench v6 | 83.8 | Score (%) | 95.7 |
| AGC-Bench - Problem Solving | 0.72 | JRT z-score | 95.1 |
| AGC-Bench - lcc_metaphor | 1.21 | Dataset z-score | 95.1 |
| AGC-Bench - simile_generation | 1.35 | Dataset z-score | 95 |
| AGC-Bench - fig_qa | 2.38 | Dataset z-score | 94.5 |
| AGC-Bench - Figurative Language | 1.02 | JRT z-score | 93.9 |
Interactive version: theaggregate.ai/model?slug=glm-4-7 · How It Works · Data refreshed daily, snapshot 2026-09-05.