Horizon - Claude Code - Hard: leaderboard
Metric: Task Pass Rate (%). Source: www.orinlabs.ai. 2 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | claude-opus-4.8 | 15.38 |
| 2 | claude-sonnet-4.5 | 4.62 |
Interactive version: theaggregate.ai/benchmark?slug=horizon-claude-code-hard · How It Works · Data refreshed daily, snapshot 2026-10-09.