Horizon - Easy — leaderboard
Metric: Task Pass Rate (%). Source: www.orinlabs.ai. 24 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | OpenClaw (LCM) - gpt-5.5 | 93.85 |
| 2 | OpenClaw (LCM) - gemini-3.5-flash | 90.77 |
| 3 | RLM - gpt-5 | 90.77 |
| 4 | RLM - claude-opus-4.8 | 87.69 |
| 5 | OpenClaw (LCM) - claude-opus-4.8 | 86.15 |
| 6 | OpenClaw (LCM) - gemini-3.1-pro-preview | 81.54 |
| 7 | RLM - claude-sonnet-4.6 | 81.54 |
| 8 | RAG - gpt-5.5 | 81.54 |
| 9 | OpenClaw (LCM) - claude-sonnet-4.5 | 78.46 |
| 10 | Codex - gpt-5-codex | 78.46 |
Interactive version: theaggregate.ai/benchmark?slug=horizon-easy · How the rankings work · Data refreshed daily, snapshot 2026-07-22.