Atlas

Benchmarks

← All benchmarks

BabyVision with Python

Multimodal · 2026-04-20

BabyVision with Python is a tool-enabled evaluation setting for BabyVision in which the model may use Python during visual reasoning. It preserves the BabyVision task set but records the tool-augmented score separately from the no-tool score.

Top models (higher is better)

ModelScore
Claude Fable 590.5
GPT-5.6 Sol88.9
Kimi K385.7
GPT-5.583.6
Opus 4.881.2
Loading Atlas data…