TerminalWorld: leaderboard
TerminalWorld Verified 200: success rate on 200 verified terminal tasks under the published Terminus 2 harness and default reasoning-effort setting. Source rows preserve solved counts, evaluation dates and model labels; this board is not an aggregate over the full 1,530-task collection.
Metric: Pass Rate (%, 200 verified tasks). Source: terminalworld.ai. Saturation forecast: Around August 2027. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.7 | 62.5 | #49 |
| 2 | Kimi K2.6 | 57.5 | #104 |
| 3 | GLM-5.1 | 57 | #97 |
| 4 | Gemini 3.1 Pro (Preview) | 55 | #58 |
| 5 | Qwen 3.6 Max Preview | 54 | #85 |
| 6 | GPT-5.5 | 53.5 | #31 |
| 7 | DeepSeek V4 Pro | 50 | #102 |
| 8 | MiniMax-M2.7 | 49 | #258 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=terminalworld · How It Works · Data refreshed daily, snapshot 2026-10-20.