Z.ai GLM-5.3 Launch - HLE with Tools: leaderboard
Metric: Accuracy (%). Source: z.ai. 8 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.6 Sol | 64.5 |
| 2 | Claude Fable 5 | 63.9 |
| 3 | GLM-5.3 | 62.5 |
| 4 | DeepSeek V4 Pro (0813) | 60 |
| 5 | Kimi K3 | 59.8 |
| 6 | Claude Opus 4.8 | 57.9 |
| 7 | Qwen 3.8 Max | 56.2 |
| 8 | GLM-5.2 | 54.7 |
Interactive version: theaggregate.ai/benchmark?slug=z-ai-glm-5-3-launch-hle-with-tools · How It Works · Data refreshed daily, snapshot 2026-09-05.