VB (Visibility Benchmark) - Second-Order Perspective: leaderboard
Metric: ToMAcc (0-1, shown times 100): accuracy on the 21 headline items of the MULTI_AGENT / SECOND_ORDER slice (what one person can see or infer about another's view), ABSTAIN counted wrong; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 9 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o | 95.2 | #333 |
| 2 | Gemini 2.5 Pro | 85.7 | #145 |
| 3 | GPT-5 | 85.7 | #91 |
| 4 | Gemini 3.1 Pro (Preview) | 81 | #54 |
| 5 | Gemma 3 12B | 71.4 | #666 |
| 6 | Claude Opus 4.5 | 66.7 | #79 |
| 7 | Claude 3.7 Sonnet | 52.4 | #241 |
| 8 | InternVL3-8B | 42.9 | #606 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=vb-visibility-benchmark-second-order-perspective · How It Works · Data refreshed daily, snapshot 2026-10-11.