PhyCritic: Multimodal Critic Models for Physical AI
This addresses the problem of evaluating and aligning AI in physical domains, but it is incremental as it builds on existing critic model frameworks.
The paper tackles the lack of reliable critic models for physical AI tasks by introducing PhyCritic, which uses a two-stage RLVR pipeline to enhance physical perception and reasoning, achieving strong performance gains over baselines on multimodal judge benchmarks.
With the rapid development of large multimodal models, reliable judge and critic models have become essential for open-ended evaluation and preference alignment, providing pairwise preferences, numerical scores, and explanatory justifications for assessing model-generated responses. However, existing critics are primarily trained in general visual domains such as captioning or image question answering, leaving physical AI tasks involving perception, causal reasoning, and planning largely underexplored. We introduce PhyCritic, a multimodal critic model optimized for physical AI through a two-stage RLVR pipeline: a physical skill warmup stage that enhances physically oriented perception and reasoning, followed by self-referential critic finetuning, where the critic generates its own prediction as an internal reference before judging candidate responses, improving judgment stability and physical correctness. Across both physical and general-purpose multimodal judge benchmarks, PhyCritic achieves strong performance gains over open-source baselines and, when applied as a policy model, further improves perception and reasoning in physically grounded tasks.