ViVerBench Release Model-based - S-Phy: leaderboard
ViVerBench3594visual verification questions,16tasks; binary true/false plus explanation. Rule-based decision correctness and model-based explanation-aware evaluation are distinct source protocols. Human/chance reference rows stay source evidence only. Metric: Verification score (0–1). Higher is better. Source: https://arxiv.org/html/2510.13804v1
Metric: Verification score (0–1). Source: arxiv.org. Saturation forecast: Around January 2027. 11 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 0.72 | #95 |
| 2 | O3 | 0.7 | #122 |
| 3 | Gemini 2.5 Pro | 0.68 | #151 |
| 4 | GPT-4o | 0.66 | #339 |
| 5 | O1 | 0.66 | #181 |
| 6 | O4 Mini | 0.57 | #175 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #7 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=viverbench-release-model-based-s-phy · How It Works · Data refreshed daily, snapshot 2026-10-17.