GraphWalks BFS: 256K to 1M Aggregate
General QA · 2025-04-12
Mean node-set F1 over all 200 problems in the GraphWalks 256K-to-1M split. Retained as a diagnostic because it overlaps the separately modeled 100-item 256K and 1M bins; its aggregate score cannot be assigned to either bin.
Top models (higher is better)
| Model | Score |
|---|---|
| Gemini 4 Argon | 84.2 |
| GPT-6 Astra | 71.8 |
| Claude Opus 5.5 | 66.8 |
| Claude Fable 5.1 | 65.0 |