RealClawBench - File Creation: leaderboard
Metric: Verifier pass rate (%) on the 57 file-creation tasks of 281 released tasks reconstructed from real OpenClaw developer-agent sessions, run in the shared OpenClaw agent runtime (file read, write, edit, search and shell tools, 600-second timeout, temperature 1.0) and scored by case-specific deterministic Python verifiers; mean of three independent runs; higher is better. Source: arxiv.org. Saturation forecast: Around December 2026. 14 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.5 | 70.2 |
| 2 | Claude Opus 4.7 | 66.1 |
| 3 | Qwen 3.6 Plus | 61.4 |
| 4 | DeepSeek V4 Flash | 60.2 |
| 5 | Kimi K2.6 | 59.1 |
| 6 | Claude Sonnet 4.6 | 56.1 |
| 7 | Gemini 3.1 Pro (Preview) | 55 |
| 8 | MiMo-V2.5-Pro | 54.4 |
| 9 | GLM-5.1 | 53.8 |
| 10 | MiniMax-M2.7 | 53.8 |
| 11 | DeepSeek V4 Pro | 52.6 |
| 12 | Claude Opus 4.6 | 46.8 |
| 13 | GPT-OSS-120B | 35.7 |
| 14 | Gemma 4 31B | 25.1 |
Interactive version: theaggregate.ai/benchmark?slug=realclawbench-file-creation · How It Works · Data refreshed daily, snapshot 2026-09-29.