PrismML Bonsai 2 27B - XHigh - Terminal-Bench 2.1 (Terminus-2 JSON): leaderboard
PrismML compares Bonsai 2 27B and Qwen3.8 27B at xhigh reasoning effort on 89 terminal tasks using the Terminus-2 agent with JSON actions. Task success (%).
Metric: Task success (%). Source: raw.githubusercontent.com. Saturation forecast: Rough model projection: around 2026. 2 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.8 27B (xHigh) (Terminus-2 JSON) | 69.7 | |
| 2 | Ternary Bonsai 2 27B (xHigh) (Terminus-2 JSON) | 52.8 |
Interactive version: theaggregate.ai/benchmark?slug=prismml-bonsai-2-27b-xhigh-terminal-bench-2-1-terminus-2-json · How It Works · Data refreshed daily, snapshot 2026-10-29.