GraphWalks BFS leaderboard
GraphWalks BFS
3 models tested · Updated 2026-04-07 · Verified sources only
Claude Mythos Preview leads at 80.0%
1
Long-context graph reasoning (256K-1M). 2x over Opus 4.6 (38.7%). GPT-5.4 scored 21.4%.
80.0%
2
BFS 256K-1M context. Well above GPT-5.4 (21.4%).
38.7%
3
BFS 256K-1M context. Well below Claude Opus (38.7%) and Claude Mythos (80.0%).
21.4%