GraphWalks Parents 256K (Anthropic): leaderboard
Anthropic GraphWalks long-context graph traversal evaluation using parent-pointer recovery tasks at 256K context.
Metric: F1 Score (%). Source: www-cdn.anthropic.com. 4 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.8 | 99.3 | #37 |
| 2 | Claude Opus 4.6 | 95.4 | #63 |
| 3 | Claude Opus 4.7 | 93.6 | #49 |
| 4 | GPT-5.5 | 90.1 | #31 |
Interactive version: theaggregate.ai/benchmark?slug=graphwalks-parents-256k-anthropic · How It Works · Data refreshed daily, snapshot 2026-10-20.