GraphWalks BFS 256K (Fable/Mythos): leaderboard
GraphWalks BFS 256K presents a text-encoded graph inside a 256K-token context and asks the model to perform breadth-first traversal over it; this slice reports Anthropic's Fable 5 / Mythos 5 system-card runs. No image input is used.
Metric: Score (%). Source: www-cdn.anthropic.com. Saturation forecast: Around January 2027. 4 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Mythos 5 | 91.1 | #2 |
| 2 | Claude Opus 4.8 | 85.9 | #37 |
| 3 | Claude Mythos Preview | 85.7 | #4 |
| 4 | GPT-5.5 | 73.7 | #31 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=graphwalks-bfs-256k-fable-mythos · How It Works · Data refreshed daily, snapshot 2026-10-20.