VPCT
- What does it measure?
- Tests intuitive physics: predicting which bucket a ball will land in after rolling down ramps.
- Who evaluates it, and how?
- 100 problems, each a picture of a ball, several ramps, and three buckets. Random guessing gets one in three, and the human volunteers all scored 100%.
- Reading the score
- Scores run from 0 to 100%. Higher is better.
Top score72.6%
Models tested10
Scoring basisHigher is betterImage-input models only
Model release date
Which AI gives the photo and chart answers people prefer most?Arena Vision
Claude Fable 5🥇
Claude Fable 5🥇Arena Vision
Claude Fable 5🥇Can it spot the wrong step from assembly photos?Furniture Assembly
Claude Fable 5.2🥇
Claude Fable 5.2🥇Furniture Assembly
Claude Fable 5.2🥇Can it draw a floor plan from a few photos of a home?Blueprint-Bench 2
GPT-6.2 Astra🥇
GPT-6.2 Astra🥇Blueprint-Bench 2
GPT-6.2 Astra🥇
Can it predict how objects will move just from a picture?
VPCT
- 🥇GoogleGemini 3 Flash72.6%
- 🥈OpenAIGPT-5Reasoning effort: High66.0%
- 🥉OpenAIGPT-5 MiniReasoning effort: Medium40.2%
- #4AnthropicClaude Opus 4.5Reasoning effort: Budget 32K40.0%
- #5AnthropicClaude Sonnet 4.5Reasoning effort: Budget 32K39.8%
- #6AnthropicClaude Opus 4Reasoning effort: Budget 16K38.0%
- #7OpenAIGPT-5 NanoReasoning effort: High37.2%
- #8AnthropicClaude Opus 4.135.0%
- #9AnthropicClaude Sonnet 4Reasoning effort: Budget 32K34.0%
- #10GoogleGemini 2.5 Flash Lite30.0%
Half of models ≤ 38.9%
The "harness" is the agent program the AI used to carry out the task. The same model can score very differently depending on its harness and reasoning effort.