HoloCount - Null-Target Prompting: leaderboard
Questions about objects absent from the image, where the correct count is zero: a subset of HoloCount (Meituan, 2026), a visual counting benchmark for multimodal LLMs. Models answer with a single integer, zero-shot; exact-match accuracy in percent, higher is better.
Metric: Accuracy (%; exact-match accuracy of the single integer count a multimodal model returns zero-shot for an image and a counting question, with a strict integer-only system prompt; questions about objects absent from the image, where the correct count is zero (250 questions); higher is better). Source: arxiv.org. Saturation forecast: Not forecast. 30 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3 VL 32B Instruct | 98 | #585 |
| 2 | Claude Sonnet 4.5 | 97.2 | #250 |
| 3 | Qwen 3 VL 8B Instruct | 96.4 | #981 |
| 4 | Claude Opus 4.7 | 95.6 | #79 |
| 5 | Qwen 2.5 VL 7B Instruct | 95.6 | #1833 |
| 6 | Kimi K2.6 | 95.6 | #178 |
| 7 | Claude Sonnet 4.6 | 95.2 | #161 |
| 8 | Qwen 2.5 VL 72B Instruct | 95.2 | #846 |
| 9 | Claude Opus 4.6 | 94.4 | #102 |
| 10 | Kimi K2.5 | 94.4 | #256 |
| 11 | Qwen 2.5 VL 32B Instruct | 93.6 | #1053 |
| 12 | Gemini 2.5 Pro | 93.4 | #266 |
| 13 | Qwen 3.5 397B A17B | 90.8 | #246 |
| 14 | Qwen 3.5 122B A10B | 90.8 | #331 |
| 15 | Claude Opus 4.8 | 90.4 | #56 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=holocount-null-target-prompting · How It Works · Data refreshed daily, snapshot 2026-11-13.