WolfBench (Terminus-2): leaderboard
Metric: Mean Success Rate (%, 89 Terminal-Bench 2.0 tasks). Source: wolfbench.ai. Saturation forecast: Rough model projection: around 2026. 40 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.6 Sol (WolfBench Run: agent=terminus-2; version=2.0.0; timeout=3600s; thinking=max; providers=openai; source-model=GPT-5.6%20Sol) | 85.17 |
| 2 | GPT-5.6 Terra (WolfBench Run: agent=terminus-2; version=2.0.0; timeout=3600s; thinking=max; providers=openai; source-model=GPT-5.6%20Terra) | 82.02 |
| 3 | GPT-5.6 Luna (WolfBench Run: agent=terminus-2; version=2.0.0; timeout=3600s; thinking=max; providers=openai; source-model=GPT-5.6%20Luna) | 76.85 |
| 4 | GPT-5.5 (WolfBench Run: agent=terminus-2; version=2.0.0; timeout=3600s; thinking=xhigh; providers=openai; source-model=GPT-5.5) | 76.78 |
| 5 | Claude Opus 4.6 (WolfBench Run: agent=terminus-2; version=2.0.0; timeout=3600s; thinking=off; providers=anthropic; source-model=Claude%20Opus%204.6) | 71.91 |
| 6 | Claude Opus 4.7 (WolfBench Run: agent=terminus-2; version=2.0.0; timeout=3600s; thinking=off; providers=anthropic; source-model=Claude%20Opus%204.7) | 71.91 |
| 7 | Gemini 3.5 Flash (WolfBench Run: agent=terminus-2; version=2.0.0; timeout=3600s; thinking=high; providers=gemini; source-model=Gemini%203.5%20Flash) | 71.46 |
| 8 | GLM-5.2 [W&B Dynamo] (WolfBench Run: agent=terminus-2; version=2.0.0; timeout=3600s; thinking=-; providers=wandb; source-model=GLM-5.2%20%5BW%26B%20Dynamo%5D) | 71.46 |
| 9 | GLM-5.2 [Z.ai] (WolfBench Run: agent=terminus-2; version=2.0.0; timeout=3600s; thinking=max; providers=openrouter; source-model=GLM-5.2%20%5BZ.ai%5D) | 71.46 |
| 10 | GPT-5.4 (WolfBench Run: agent=terminus-2; version=2.0.0; timeout=3600s; thinking=xhigh; providers=openai; source-model=GPT-5.4) | 71.16 |
Interactive version: theaggregate.ai/benchmark?slug=wolfbench-terminus-2 · How It Works · Data refreshed daily, snapshot 2026-10-09.