Google Launches EmbeddingGemma 2 and Reflection Debuts Beam as New Open-Source Models
Summary
Google releases EmbeddingGemma 2, an open-source 740-million-parameter on-device model that handles text, code, images, video and audio, and which Google claims is best-in-class for embedded models under 1 billion parameters. New York startup Reflection launches Beam, a 500-billion-parameter open-source model for coding, reasoning and agentic work. Reflection says Beam runs frontier reasoning up to four times cheaper than GLM-5.2.
Key Points
- Google launches EmbeddingGemma 2, an open-source on-device model with 740 million parameters that handles text, code, images, video and audio.
- Google says the multilingual model is best-in-class for embedded models under 1 billion parameters, finding voice memo clips and searching hours of audio from a text request.
- New York startup Reflection launches Beam, an open-source 500 billion parameter model it says offers frontier reasoning up to four times cheaper than GLM-5.2.