Model updates

Official news, first rumor reports and substantive progress. Each entry keeps the information published at that time.

• mlx-community publishes EmbeddingGemma 2 MLX builds

  The mlx-community organization on Hugging Face published MLX builds of Google's EmbeddingGemma 2 embedding model, including an NVFP4 quantization.

  UnconfirmedPublished newsOct 6Source: Hugging Face @mlx-communityPart of EmbeddingGemma 2
• Google DeepMind releases EmbeddingGemma 2, a natively multimodal open embedding model

  Google DeepMind released EmbeddingGemma 2, an open multimodal embedding model under Apache 2.0 that maps text including code, images, video, and audio into a single 768-dimensional vector space. The model has 740 million total parameters, combining a 270M text model with 170M vision and 300M audio encoders. It is designed to run on consumer hardware like phones and laptops for low-latency search, RAG, classification, and clustering. EmbeddingGemma 2 builds on the Gemma 4 architecture.

  OfficialOfficial updateOct 6Source: Google DeepMindPart of EmbeddingGemma 2

RSS for this view →