Hugging Face

HuggingFace has released a new version of their SmolLM series alongside a very detailed writeup of all the decisions made, including details about all the used datasets. Similar to other models, it supports a thinking and non-thinking mode. We played around with it and were impressed by the quality, it really is a model on the level of the small Qwen3 models and comes with all artifacts released - one of the few "open-source" models today.

HuggingFace has updated their SmolVLM models, focusing on video understanding. They also released an iPhone app to run the 500M model natively on iOS.

Alongside the new SmolLM series, HuggingFace also released small vision models based on the Idefics architecture.

New models in the SmolLLM series.

HuggingFace released some decent small language models (135M, 350M, and this). Great to have more attention here. They also released a very fun demo that runs in your browser.