Long Context Benchmark

GraphWalks (BFS, 256K)

GraphWalks BFS mode at 256K context over the openai/graphwalks dataset (official dataset page). Score as reported on the Inkling 2 Mini card.

Category

Long Context

Score type

percent

Scale

100%

higher is better

Models tested

2

What is GraphWalks (BFS, 256K)?

GraphWalks (BFS, 256K) is a benchmark for reading and reasoning over very long inputs. The official description says: "GraphWalks BFS mode at 256K context over the openai/graphwalks dataset (official dataset page)". Below we describe in simple words what benchmarks of this type generally examine.

What does it test?

GraphWalks (BFS, 256K) measures reading and reasoning over very long inputs. In detail, it checks:

  • finding relevant facts buried in documents hundreds of thousands of tokens long
  • retrieving multiple needles hidden across the full context
  • reasoning over whole books, codebases or conversation histories
  • staying accurate near the context window limit (no 'lost in the middle' falloff)

How to read the score

The score is accuracy: the percentage of tasks answered or solved correctly. 100% means every task was solved perfectly; higher is better.

Official methodology note: GraphWalks BFS mode at 256K context over the openai/graphwalks dataset (official dataset page). Score as reported on the Inkling 2 Mini card.

Frequently asked questions

What is the GraphWalks (BFS, 256K) benchmark?

GraphWalks (BFS, 256K) is a benchmark for reading and reasoning over very long inputs. The official description says: 'GraphWalks BFS mode at 256K context over the openai/graphwalks dataset (official dataset page)'. Below we describe in simple words what benchmarks of this type generally examine.

What does GraphWalks (BFS, 256K) test?

GraphWalks (BFS, 256K) mainly measures reading and reasoning over very long inputs. Concretely, it checks: finding relevant facts buried in documents hundreds of thousands of tokens long; retrieving multiple needles hidden across the full context; reasoning over whole books, codebases or conversation histories.

How is the GraphWalks (BFS, 256K) score calculated?

The score is accuracy: the percentage of tasks answered or solved correctly. 100% means every task was solved perfectly; higher is better. The declared maximum score of this benchmark is 100.0.

Where do I get GraphWalks (BFS, 256K) data and scores?

The official source is https://huggingface.co/datasets/openai/graphwalks. The official benchmark page/reference is linked on this page. Every score on Open LLM Radar is also available as open data: download the scores JSON (http://open-llm-radar-api:8000/api/v1/benchmarks/graphwalks-bfs-256k/scores) or pull it directly from the Open LLM Radar API with any HTTP client.

Model scores 2 models tested

Reset
# Model Score Date Source Verified
1 Inkling 2 Mini
Thinking Machines
69.7%
20.11.2026 Report Verified
2 Qwen3.8-27B
Qwen Team (Alibaba Cloud)
50.4%
20.11.2026 Report Unverified

What now?

You've seen how models score on GraphWalks (BFS, 256K) — here's what to do next.