KOFFVQA - Relationship: leaderboard
KOFFVQA subscore for spatial and semantic relationships between visual entities.
Metric: Score (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 82 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | O1 (2024-12-17) | 86.67 | #149 |
| 2 | GPT-4.1 | 83.33 | #248 |
| 3 | O4 Mini (2025-04-16) | 82 | #179 |
| 4 | GPT-4o (2024-11-20) | 80 | #381 |
| 5 | Gemini 2.5 Pro (Preview 03-25) | 79.67 | #123 |
| 6 | GPT-4.1 Mini | 79 | #355 |
| 7 | O3 (2025-04-16) | 78 | #121 |
| 8 | Gemini 2.5 Pro (Preview 05-06) | 77.67 | #147 |
| 9 | GPT-4.5 (Preview) | 76.67 | #244 |
| 10 | GPT-4 Turbo | 76.33 | #404 |
| 11 | Llama 4 Maverick Instruct | 74.33 | #456 |
| 12 | Gemini 2.5 Flash (Preview 05-20) | 73.33 | #216 |
| 13 | InternVL3-38B | 70.67 | #405 |
| 14 | Gemini 2.5 Flash (Preview 04-17) | 70.67 | #237 |
| 15 | Gemini 1.5 Pro (002) | 69.33 | #349 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=koffvqa-relationship · How It Works · Data refreshed daily, snapshot 2026-10-20.