VPCT uses 100 simulator-generated image problems to measure whether vision-language models can predict which bucket a ball will enter from the initial physical setup.
Browse the latest scores, model modes, release dates, and parameter sizes for VPCT.
Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology
| Rank | Model | License | |||
|---|---|---|---|---|---|
 Gemini 3 Pro Standard Mode | 91.00 | 2026-02-01 | Unknown | Closed | |
 GPT-5.2 Thinking Level · Extra High | 84.00 | 2025-12-11 | Unknown | Closed | |
Gemini 3.0 Flash Standard Mode | 72.60 | 2025-12-17 | Unknown | Closed | |
4 | GPT-5.2 Thinking Level · High | 67.00 | 2025-12-11 | Unknown | Closed |
5 | GPT-5 Thinking Level · High | 66.00 | 2025-08-07 | Unknown | Closed |
6 | GPT-5 Thinking Level · Medium | 63.20 | 2025-08-07 | Unknown | Closed |
7 | GPT-5.1 Thinking Level · High | 58.70 | 2025-11-12 | Unknown | Closed |
8 | OpenAI o4 - mini Thinking Level · Medium | 57.50 | 2025-04-16 | Unknown | Closed |
9 | GPT-5.1 Thinking Level · Medium | 53.30 | 2025-11-12 | Unknown | Closed |
10 | OpenAI o3 Thinking Level · Medium | 52.00 | 2025-04-16 | Unknown | Closed |
11 | Gemini 2.5 Pro Experimental 03-25 Standard Mode | 48.00 | 2025-03-25 | Unknown | Closed |
12 | Gemini 2.5-Pro Standard Mode | 46.40 | 2025-06-05 | Unknown | Closed |
13 | Gemini-2.5-Pro-Preview-05-06 Standard Mode | 40.50 | 2025-05-06 | Unknown | Closed |
14 | GPT-5-mini Thinking Level · Medium | 40.20 | 2025-08-07 | Unknown | Closed |
15 | GPT-4o(2024-11-20) Standard Mode | 40.00 | 2024-11-20 | Unknown | Closed |
16 |  Opus 4.5 32K | 40.00 | 2025-11-25 | Unknown | Closed |
17 | 39.80 | 2025-09-30 | Unknown | Closed | |
18 | Claude Sonnet 3.7 Standard Mode | 39.00 | 2025-02-25 | Unknown | Closed |
19 | GPT-5-mini Thinking Level · High | 39.00 | 2025-08-07 | Unknown | Closed |
20 | Claude Sonnet 4.5 Standard Mode | 38.00 | 2025-09-30 | Unknown | Closed |
21 | OpenAI o1 Thinking Level · Medium | 37.00 | 2024-12-05 | Unknown | Closed |
22 | 35.00 | 2025-02-25 | Unknown | Closed | |
23 | GPT-4o mini Standard Mode | 34.00 | 2024-07-18 | Unknown | Closed |
24 | Claude 3.5 Sonnet New Standard Mode | 33.00 | 2024-10-22 | Unknown | Closed |
In the official leaderboard snapshot observed on 2026-08-12, Gemini 3 Pro Preview scored 91.0% accuracy.
