PerceptionBench - Counting — leaderboard
Metric: Counting (%). Source: huggingface.co. 16 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.6 Sol | 62.4 |
| 2 | Kimi K3 | 59.7 |
| 3 | Gemini 3.1 Pro (Preview) | 56.9 |
| 4 | GPT-5.5 | 55.8 |
| 5 | Qwen 3.7 Plus | 53.3 |
| 6 | Claude Fable 5 | 52.9 |
| 7 | Seed 2.1 Pro | 51.2 |
| 8 | Qwen 3.5 397B A17B | 49.1 |
| 9 | Kimi K2.6 | 45.2 |
| 10 | Claude Opus 4.8 | 44.2 |
| 11 | Gemini 3.5 Flash | 43.6 |
| 12 | GLM-5V Turbo | 40 |
| 13 | Grok 4.5 | 35.2 |
| 14 | Gemma 4 31B | 33.9 |
| 15 | GLM-4.6V | 31.8 |
Interactive version: theaggregate.ai/benchmark?slug=perceptionbench-counting · How It Works · Data refreshed daily, snapshot 2026-08-05.