GraphWalks BFSBrowse 284 families
GraphWalks BFS
Can the model retain and reason over evidence spread across a very long input?
Version / measure
- Domain
- Long Context
- Ranked models
- 7
- Reported values
- 7
- Reports
- 1 · lab
- Direction
- Higher is better
- Status
- Version not specified
Rankings
Best reported score per model across effort settings
| # | Model | Best setting | Score |
|---|---|---|---|
| 1 | Claude Mythos 5Anthropic | reportedcost n/a1 value · Jul 9 · source | 79.4% |
| 2 | GPT-5.6 SolOpenAI | reportedcost n/a1 value · Jul 9 · source | 77.1% |
| 3 | Claude Mythos PreviewAnthropic | reportedcost n/a1 value · Jul 9 · source | 74.3% |
| 4 | GPT-5.6 TerraOpenAI | reportedcost n/a1 value · Jul 9 · source | 71.2% |
| 5 | Claude Opus 4.8Anthropic | reportedcost n/a1 value · Jul 9 · source | 68.1% |
| 6 | GPT-5.6 LunaOpenAI | reportedcost n/a1 value · Jul 9 · source | 51.2% |
| 7 | GPT-5.5OpenAI | reportedcost n/a1 value · Jul 9 · source | 45.4% |
Effort curve
Score vs. reported cost. Lines connect complete sweeps only.
No cost-linked effort sweep for this version.
Definition and comparison boundarypublic methodology
Can the model retain and reason over evidence spread across a very long input?
Long-context retrieval and reasoning, not context-window size alone. Higher is better. The value is the percentage reported in this lab's table.
Harness, tools, prompts, attempt count, and benchmark version can materially change the result. Compare within one reporting context.
Method / source ↗