Z.ai GLM-5.3 Launch - Terminal Bench 3.0: leaderboard
Terminal-Bench 3.0 terminal-agent tasks as reported in Z.ai's GLM-5.3 launch table, using Claude Code with maximum reasoning effort, isolated task containers, official verifiers, and avg@3 over three rollouts; Z.ai-run frontier comparison.
Metric: Accuracy (%). Source: z.ai. Saturation forecast: Around September 2027. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.6 Sol | 34.6 | #19 |
| 2 | Claude Fable 5 | 33.7 | #16 |
| 3 | GLM-5.3 | 28.3 | #39 |
| 4 | Claude Opus 4.8 | 21.1 | #37 |
| 5 | Kimi K3 | 17.4 | #30 |
| 6 | GLM-5.2 | 4.6 | #73 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=z-ai-glm-5-3-launch-terminal-bench-3-0 · How It Works · Data refreshed daily, snapshot 2026-10-20.