Google Debuts Gemma 4 Open-Weight Models
Google DeepMind expanded its Gemma line with open-weight models ranging from 2B and 4B edge versions to 26B MoE and 31B dense versions. The release is notable because it uses Gemini 3 technology and is the first Gemma release under Apache 2.0. It also supports 256K context and multimodal processing.
- Gemma 4 now spans edge and larger deployments, from 2B and 4B models to 26B MoE and 31B dense versions.
- The lineup is built on Gemini 3 technology and released under Apache 2.0 for the first time.
- Google says the models support 256K context and multimodal processing.
