Multimodal benchmark # VPCT leaderboard > VPCT results for 24 AI models, led by Gemini 3 Pro at 91%. What the benchmark measures, who runs it, and a source for every score. - Canonical page: https://noometry.com/benchmarks/vpct - Last updated: 2026-10-10 - Title: VPCT Leaderboard (October 2026): Scores by Model | Noometry As of October 2026, Gemini 3 Pro has the highest published VPCT score on Noometry at 91%, out of 24 models with results. Last verified October 10, 2026 ## About VPCT Visual physics comprehension: predict where a ball will land in a diagram. - **Category:** [Multimodal](https://noometry.com/best/multimodal) - **Introduced:** 2025 - **Format:** Multiple choice - **Unit:** Percent (random guessing ≈ 33.3%) - **Official site:** [cbrower.dev](https://cbrower.dev/vpct) ## Top 15 models Top models on VPCT 1. Gemini 3 Pro 91% 2. GPT-5.2 84% 3. Gemini 3 Flash Preview 72.6% 4. GPT-5 66% 5. GPT-5.1 58.7% 6. o4-mini 57.5% 7. o3 52% 8. Gemini 2.5 Pro 48% 9. Gemini 2.5 Flash 46.2% 10. GPT-4.5 45% 11. GPT-5 Mini 40.2% 12. Claude Opus 4.5 40% 13. GPT-4o 40% 14. Claude Sonnet 4.5 39.8% 15. Claude 3.7 Sonnet 39% 16. 050100 Sponsored placements are available on pages like this one. [Advertise on Noometry](https://noometry.com/advertise) ## All results VPCT results by model | # | Model | Provider | Score | Setting | Source | Date | | --- | --- | --- | --- | --- | --- | --- | | 1 | [Gemini 3 Pro](https://noometry.com/models/gemini-3-pro) | [![](/logos/google.svg) Google](https://noometry.com/providers/google) | 91% | | [Epoch AI](https://epoch.ai/benchmarks) | | | 2 | [GPT-5.2](https://noometry.com/models/gpt-5-2) | [OpenAI](https://noometry.com/providers/openai) | 84% | xhigh | [Epoch AI](https://epoch.ai/benchmarks) | | | 3 | [Gemini 3 Flash Preview](https://noometry.com/models/gemini-3-flash-preview) | [![](/logos/google.svg) Google](https://noometry.com/providers/google) | 72.6% | | [Epoch AI](https://epoch.ai/benchmarks) | | | 4 | [GPT-5](https://noometry.com/models/gpt-5) | [OpenAI](https://noometry.com/providers/openai) | 66% | high | [Epoch AI](https://epoch.ai/benchmarks) | | | 5 | [GPT-5.1](https://noometry.com/models/gpt-5-1) | [OpenAI](https://noometry.com/providers/openai) | 58.7% | high | [Epoch AI](https://epoch.ai/benchmarks) | | | 6 | [o4-mini](https://noometry.com/models/o4-mini) | [OpenAI](https://noometry.com/providers/openai) | 57.5% | medium | [Epoch AI](https://epoch.ai/benchmarks) | | | 7 | [o3](https://noometry.com/models/o3) | [OpenAI](https://noometry.com/providers/openai) | 52% | medium | [Epoch AI](https://epoch.ai/benchmarks) | | | 8 | [Gemini 2.5 Pro](https://noometry.com/models/gemini-2-5-pro) | [![](/logos/google.svg) Google](https://noometry.com/providers/google) | 48% | | [Epoch AI](https://epoch.ai/benchmarks) | | | 9 | [Gemini 2.5 Flash](https://noometry.com/models/gemini-2-5-flash) | [![](/logos/google.svg) Google](https://noometry.com/providers/google) | 46.2% | | [Epoch AI](https://epoch.ai/benchmarks) | | | 10 | [GPT-4.5](https://noometry.com/models/gpt-4-5) | [OpenAI](https://noometry.com/providers/openai) | 45% | | [Epoch AI](https://epoch.ai/benchmarks) | | | 11 | [GPT-5 Mini](https://noometry.com/models/gpt-5-mini) | [OpenAI](https://noometry.com/providers/openai) | 40.2% | medium | [Epoch AI](https://epoch.ai/benchmarks) | | | 12 | [Claude Opus 4.5](https://noometry.com/models/claude-opus-4-5) | [Anthropic](https://noometry.com/providers/anthropic) | 40% | 32K | [Epoch AI](https://epoch.ai/benchmarks) | | | 13 | [GPT-4o](https://noometry.com/models/gpt-4o) | [OpenAI](https://noometry.com/providers/openai) | 40% | | [Epoch AI](https://epoch.ai/benchmarks) | | | 14 | [Claude Sonnet 4.5](https://noometry.com/models/claude-sonnet-4-5) | [Anthropic](https://noometry.com/providers/anthropic) | 39.8% | 32K | [Epoch AI](https://epoch.ai/benchmarks) | | | 15 | [Claude 3.7 Sonnet](https://noometry.com/models/claude-3-7-sonnet) | [Anthropic](https://noometry.com/providers/anthropic) | 39% | | [Epoch AI](https://epoch.ai/benchmarks) | | | 16 | [Claude Opus 4](https://noometry.com/models/claude-opus-4) | [Anthropic](https://noometry.com/providers/anthropic) | 38% | 16K | [Epoch AI](https://epoch.ai/benchmarks) | | | 17 | [GPT-5 Nano](https://noometry.com/models/gpt-5-nano) | [OpenAI](https://noometry.com/providers/openai) | 37.2% | high | [Epoch AI](https://epoch.ai/benchmarks) | | | 18 | [o1](https://noometry.com/models/o1) | [OpenAI](https://noometry.com/providers/openai) | 37% | medium | [Epoch AI](https://epoch.ai/benchmarks) | | | 19 | [Claude Opus 4.1](https://noometry.com/models/claude-opus-4-1) | [Anthropic](https://noometry.com/providers/anthropic) | 35% | | [Epoch AI](https://epoch.ai/benchmarks) | | | 20 | [Claude Sonnet 4](https://noometry.com/models/claude-sonnet-4) | [Anthropic](https://noometry.com/providers/anthropic) | 34% | 32K | [Epoch AI](https://epoch.ai/benchmarks) | | | 21 | [GPT-4o mini](https://noometry.com/models/gpt-4o-mini) | [OpenAI](https://noometry.com/providers/openai) | 34% | | [Epoch AI](https://epoch.ai/benchmarks) | | | 22 | [Claude 3.5 Sonnet](https://noometry.com/models/claude-3-5-sonnet) | [Anthropic](https://noometry.com/providers/anthropic) | 33% | | [Epoch AI](https://epoch.ai/benchmarks) | | | 23 | [Claude 3.5 Sonnet](https://noometry.com/models/claude-3-5-sonnet) | [Anthropic](https://noometry.com/providers/anthropic) | 33% | | [Epoch AI](https://epoch.ai/benchmarks) | | | 24 | [Gemini 2.5 Flash-Lite](https://noometry.com/models/gemini-2-5-flash-lite) | [![](/logos/google.svg) Google](https://noometry.com/providers/google) | 30% | | [Epoch AI](https://epoch.ai/benchmarks) | | ## Compare the leaders - [Gemini 3 Pro vs GPT-5.2](https://noometry.com/compare/gemini-3-pro-vs-gpt-5-2) - [Gemini 3 Pro vs Gemini 3 Flash Preview](https://noometry.com/compare/gemini-3-flash-preview-vs-gemini-3-pro) - [Gemini 3 Pro vs GPT-5](https://noometry.com/compare/gemini-3-pro-vs-gpt-5) - [Gemini 3 Pro vs GPT-5.1](https://noometry.com/compare/gemini-3-pro-vs-gpt-5-1) - [GPT-5.2 vs Gemini 3 Flash Preview](https://noometry.com/compare/gemini-3-flash-preview-vs-gpt-5-2) - [GPT-5.2 vs GPT-5](https://noometry.com/compare/gpt-5-vs-gpt-5-2) ## Other multimodal benchmarks - [LMArena Vision](https://noometry.com/benchmarks/arena-vision) - [Video-MME](https://noometry.com/benchmarks/video-mme) - [GeoBench](https://noometry.com/benchmarks/geobench) - [Blueprint-Bench 2](https://noometry.com/benchmarks/blueprint-bench-2) - [Furniture Assembly](https://noometry.com/benchmarks/furniture-assembly) - [LMArena Document](https://noometry.com/benchmarks/arena-document) (reference) - [MindCube](https://noometry.com/benchmarks/mindcube) (reference) - [ScienceQA](https://noometry.com/benchmarks/scienceqa) (reference) - [SpatialViz-Bench](https://noometry.com/benchmarks/spatialviz-bench) (reference) ## Frequently asked questions ### What does VPCT measure? Visual physics comprehension: predict where a ball will land in a diagram. ### Which model has the highest VPCT score? As of October 2026, Gemini 3 Pro has the highest published VPCT score on Noometry at 91%, out of 24 models with results. ### Cite this page Noometry. (2026). VPCT leaderboard. Retrieved October 10, 2026, from https://noometry.com/benchmarks/vpct Quote Noometry with a link back to this page. It is also available in [Markdown](https://noometry.com/md/benchmarks/vpct.md).