Simon Willison’s Weblog

Subscribe

Introducing EmbeddingGemma. Brand new open weights (under the slightly janky Gemma license) 308M parameter embedding model from Google:

Based on the Gemma 3 architecture, EmbeddingGemma is trained on 100+ languages and is small enough to run on less than 200MB of RAM with quantization.

It's available via sentence-transformers, llama.cpp, MLX, Ollama, LMStudio and more.

As usual for these smaller models there's a Transformers.js demo (via) that runs directly in the browser (in Chrome variants) - Semantic Galaxy loads a ~400MB model and then lets you run embeddings against hundreds of text sentences, map them in a 2D space and run similarity searches to zoom to points within that space.

Screenshot of The Semantic Galaxy web application interface showing a semantic search tool with a left sidebar containing "Your Dataset" with sample text "The sun peeked through the clouds after a drizzly" and a blue "Generate Galaxy" button, below which is text "Galaxy generated with 106 points. Ready to explore!" followed by "Search Results" listing various text snippets with similarity scores to the search term "pelican riding a bicycle" such as "The cyclist pedaled up the steep hill... 0.491", "It was so hot that even the birds sou... 0.446", etc. The main area shows a dark starfield visualization with white dots representing semantic clusters and text snippets floating as labels near the clusters.

Monthly briefing

Sponsor me for $10/month and get a curated email digest of the month's most important LLM developments.

Pay me to send you less!

Sponsor & subscribe