Model vs Harness - Terminal-Bench 2 Mean-of-4: leaderboard
Model vs Harness v1, four matched repetitions, all scheduled attempts included even if runtime fails; official task correctness differs from valid-completion rate. Harnesses: DeepAgents Code 0.1.3, OpenCode 1.15.6, Pi CLI 0.79.3, Codex CLI (unreported version). DeepSeek V4 Flash Preview official API; MiniMax M3, Kimi K2.7 Code, DeepSeek V4 Pro 0813 local SGLang on eight H200 GPUs. Qwen3.6-35B-A3B explicit HF checkpoint. Fixed GPT-5.5 xhigh native Codex subscription and DeepSeek V4 Pro 0813 native Harness SDK 0.1.0rc6 are complete-system references. No unreported reasoning effort inferred. All 89 tasks through Harbor with native verifiers/time limits. Mean-of-4. Metric: Success rate (%).
Metric: Success rate (%). Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 18 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Model vs Harness GPT-5.5 (xhigh) (Harness: Codex CLI native) | 60.39 | |
| 2 | Model vs Harness DeepSeek-V4-Pro-0813 (Harness: DeepSeek Harness SDK 0.1.0rc6) | 46.91 | |
| 3 | Model vs Harness DeepSeek-V4-Flash-Preview (Harness: DeepAgents 0.1.3) | 45.79 | |
| 4 | Model vs Harness DeepSeek-V4-Flash-Preview (Harness: Pi CLI 0.79.3) | 44.66 | |
| 5 | Model vs Harness Kimi-K2.7-Code (Harness: Pi CLI 0.79.3) | 43.82 | |
| 6 | Model vs Harness DeepSeek-V4-Flash-Preview (Harness: OpenCode 1.15.6) | 40.45 | |
| 7 | Model vs Harness MiniMax-M3 (Harness: OpenCode 1.15.6) | 39.61 | |
| 8 | Model vs Harness Kimi-K2.7-Code (Harness: OpenCode 1.15.6) | 39.04 | |
| 9 | Model vs Harness DeepSeek-V4-Flash-Preview (Harness: Codex CLI unspecified version) | 37.64 | |
| 10 | Model vs Harness MiniMax-M3 (Harness: Codex CLI unspecified version) | 34.55 |
Interactive version: theaggregate.ai/benchmark?slug=model-vs-harness-terminal-bench-2-mean-of-4 · How It Works · Data refreshed daily, snapshot 2026-10-27.