Z.ai GLM-5.3 Launch - ExploitGym (2h): leaderboard
Metric: Successful Intended Exploits (#). Source: z.ai. 7 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.6 Sol | 216 |
| 2 | Claude Fable 5 | 181 |
| 3 | GLM-5.3 | 105 |
| 4 | Claude Opus 4.8 | 80 |
| 5 | Kimi K3 | 36 |
| 6 | GLM-5.2 | 29 |
| 7 | Qwen 3.8 Max | 14 |
Interactive version: theaggregate.ai/benchmark?slug=z-ai-glm-5-3-launch-exploitgym-2h · How It Works · Data refreshed daily, snapshot 2026-09-05.