Graphwalks BFS <128k: leaderboard
Graphwalks breadth-first-search task below 128k context, testing graph traversal and reachable-node reasoning from serialized graph inputs.
Metric: Score (%). Source: huggingface.co. Source: llm-stats.com. Saturation forecast: Estimated already saturated. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 4 Argon | 99.7 | #8 |
| 2 | GPT-5.2 | 94 | #110 |
| 3 | GPT-5.4 | 93 | #79 |
| 4 | GPT-5 | 78.3 | #98 |
| 5 | GPT-5.4 Mini | 76.3 | #212 |
| 6 | GPT-5.4 Nano | 73.4 | #328 |
| 7 | GPT-4.5 | 72.3 | #274 |
| 8 | Seed 2.0 Mini | 64.1 | #276 |
| 9 | GPT-4.1 Mini | 61.7 | #355 |
| 10 | GPT-4.1 | 61.7 | #248 |
| 11 | O3 Mini | 51 | #272 |
| 12 | GPT-4o | 41.7 | #343 |
| 13 | GPT-4.1 Nano | 25 | #731 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=graphwalks-bfs-under-128k · How It Works · Data refreshed daily, snapshot 2026-10-20.