HOI-Edit - Foundational Editing: leaderboard

Metric: Interaction score (0-100) on the 357 Level-1 edits (create, remove or change one human-object interaction), scored by HOI-Eval: the human and object boxes are tracked into the edited image and Gemini 2.5 Pro rates the interaction from tagged crops; higher is better. Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 8 models tracked.

Top models

#ModelScore
1Nano Banana72.71
2Wan2.2-I2V-A14B69.08
3Bagel63.26
4Qwen-Image-Edit PLUS61.28
5ChronoEdit58.23
6Step1X-Edit53.96
7FLUX.1 Kontext [dev]49.61
8ByteMorph44.69

Interactive version: theaggregate.ai/benchmark?slug=hoi-edit-foundational-editing · How It Works · Data refreshed daily, snapshot 2026-09-29.