Doc2Harness - Interactive terminal: leaderboard
Table 2; three independent builds per model in the same fixed builder harness; development document plus interface skeleton; 328 end-to-end test cases; frozen artifact tested by scripted mock model; category means and overall weighted by test count. Metric: Pass rate (%). Higher is better. Source: https://arxiv.org/html/2610.20344v1
Metric: Pass rate (%). Source: arxiv.org. Saturation forecast: Around January 2027. 19 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.8 Max (0902) | 93.3 | #28 |
| 2 | Claude Opus 5 | 91.7 | #12 |
| 3 | GPT-6 Astra | 91.7 | #5 |
| 4 | Kimi K3 | 86.7 | #30 |
| 5 | Claude Opus 4.7 | 85 | #49 |
| 6 | GLM-5.3 | 85 | #40 |
| 7 | Claude Opus 4.8 | 83.3 | #38 |
| 8 | GPT-5.6 Sol | 83.3 | #18 |
| 9 | DeepSeek V4.1 Flash | 81.7 | #36 |
| 10 | DeepSeek V4 Pro | 78.3 | #99 |
| 11 | Grok 4.7 | 78.3 | #27 |
| 12 | Hy4 preview | 78.3 | #39 |
| 13 | GPT-5.5 | 76.7 | #32 |
| 14 | MiMo-V2.6-Flash | 76.7 | #62 |
| 15 | Step 5 Preview | 76.7 | #41 |
No result here: #3 Claude Opus 5.5, #7 Gemini 4 Argon, #8 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=doc2harness-interactive-terminal · How It Works · Data refreshed daily, snapshot 2026-10-16.