LLM Stats (V*): leaderboard
V* benchmark testing fine-grained visual grounding — locating small or occluded objects in high-resolution images that require focused attention.
Metric: Score (%). Source: llm-stats.com. Saturation forecast: Estimated already saturated. 7 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Kimi K2.6 | 96.9 | #104 |
| 2 | Qwen 3.6 Plus | 96.9 | #120 |
| 3 | Qwen 3.6 27B | 94.7 | #170 |
| 4 | Qwen 3.5 27B | 93.7 | #187 |
| 5 | Qwen 3.5 122B A10B | 93.2 | #173 |
| 6 | Qwen 3.5 35B A3B | 92.7 | #256 |
| 7 | GLM-5V Turbo | 89 | #166 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=llm-stats-v-star · How It Works · Data refreshed daily, snapshot 2026-10-20.