LoopsBench - Claude Code Loop - Dependency Depth: leaderboard
Metric: Dependency depth reached before the first residual handoff (0-1). Source: arxiv.org. 8 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.5 | 0.47 |
| 2 | DeepSeek V4 Pro | 0.46 |
| 3 | GLM-5.1 | 0.44 |
| 4 | Gemini 3.1 Pro (Preview) | 0.31 |
| 5 | Grok 4.1 Fast (Reasoning) | 0.15 |
| 6 | Kimi K2.6 | 0.14 |
Interactive version: theaggregate.ai/benchmark?slug=loopsbench-claude-code-loop-dependency-depth · How It Works · Data refreshed daily, snapshot 2026-09-19.