Skip to content

BabyVision with Python

A Python-assisted BabyVision evaluation for fine-grained visual perception and grounded reasoning.

5 models

Higher is better.

  1. 1DeepSeek V4.1 Flash 1017DeepSeek92.5
  2. 2Qwen3.8 MaxAlibaba91.3
  3. 3MiniMax M3.1 FlashMiniMax91.0
  4. 4Kimi K3Moonshot AI85.7
  5. 5Qwen3.8-27BAlibaba85.6
Higher is better. Bar colour: blue proprietary, orange open weights, green other.

About this benchmark

Metric
Tool-augmented multimodal score
Tasks
Visual perception tasks with Python
Authors
Moonshot AI (2026)

Scores are as reported by model developers or evaluators and may use different settings (for example reasoning effort or tools). Compare with care.