reasoning benchmark
Graphwalks BFS >128k
A graph reasoning benchmark that evaluates language models' ability to perform breadth-first search (BFS) operations on graphs with context length over 128k tokens, testing long-context reasoning capabilities.
Updated Aug 7, 2026
Published models11
Registry coverage11
MetricScore
EvidenceB
Graphwalks BFS >128k leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 3 | OP | 90.7% | 90.5% | 22 | C | |
| 4 | OP | 81.3% | 85.7% | 22 | C | |
| 5 | AN | 80.0% | 81.0% | 22 | C | |
| 7 | OP | 76.9% | 71.4% | 22 | C | |
| 11 | AN | 68.1% | 52.4% | 22 | C | |
| 14 | AN | 61.5% | 38.1% | 22 | C | |
| 16 | OP | 45.4% | 28.6% | 22 | C | |
| 19 | OP | 21.4% | 14.3% | 22 | C | |
| 20 | OP | 19.0% | 9.5% | 22 | C | |
| 21 | OP | 15.0% | 4.8% | 22 | C | |
| 22 | OP | 2.9% | 0.0% | 22 | C |
Score distribution
Top published rows on the benchmark's original scale.
Graphwalks BFS >128k highlights
The top published results on this benchmark's own scale.
What is Graphwalks BFS >128k?
Definition and scoring fields from the benchmark registry.
A graph reasoning benchmark that evaluates language models' ability to perform breadth-first search (BFS) operations on graphs with context length over 128k tokens, testing long-context reasoning capabilities.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- Graphwalks BFS >128k
- Modality
- text
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- graphwalks-bfs->128k|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about Graphwalks BFS >128k.
Which model scores highest on Graphwalks BFS >128k?
GPT-5.6 Sol is currently ranked first with 90.7%.
What does Graphwalks BFS >128k measure?
A graph reasoning benchmark that evaluates language models' ability to perform breadth-first search (BFS) operations on graphs with context length over 128k tokens, testing long-context reasoning capabilities.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
11 unique published model results are currently shown.
Does this benchmark affect the overall score?
This benchmark is marked as eligible for the current LLMBoard capability methodology.