SmolVLM Grows Smaller – Introducing the 256M & 500M Models!

In three linesHugging Face releases SmolVLM 256M and 500M, ultra-compact vision-language models. These variants drastically reduce model size while maintaining multimodal capabilities, targeting edge deployments and resource-constrained environments.

Summary generated by Claude — human-verified