ReaLM - FigStep: leaderboard

Metric: Attack success rate (%; share of adversarial cases where the model gives an incorrect, task-relevant answer) under FigStep (misleading text rendered as an auxiliary image) on the 832 ReaLM physical-world cases (driving, manipulation, grasping, physics, egocentric, diagnostics and scene QA; image, question and ground-truth answer), black-box one-shot evaluation at temperature 0, a Qwen3-8B judge extracting the answer; lower is better. Source: arxiv.org. Saturation forecast: Around 2030. 13 models tracked.

Top models

#ModelScore
1Gemini 3 Flash43.6
2GPT-4.1 Mini45.4
3Qwen 3.5 9B47
4Seed 2.0 Mini47.2
5Kimi K2.548.3
6Qwen 3.6 Flash49.2
7Qwen 3.5 122B A10B49.3
8Qwen 3 VL 32B49.3
9Qwen 3.5 27B49.5

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