OpenAI has introduced sCM, a new approach to continuous-time consistency models that significantly speeds up generative AI sampling. This method simplifies and stabilizes training, allowing models to generate high-quality samples comparable to diffusion models in just two steps, a ~50x improvement in speed. The research demonstrates that sCMs scale effectively with teacher diffusion models and can achieve state-of-the-art results on benchmarks like ImageNet with reduced computational cost. AI
Summary written by gemini-2.5-flash-lite from 4 sources. How we write summaries →
RANK_REASON OpenAI published research on a new class of generative models called consistency models, detailing improvements and scaling techniques.