AI Development
Google releases Gemma 4 open models with multimodal support and Apache 2.0 license
Editorial Analysis
Google DeepMind launched Gemma 4 family of open models (E2B, E4B, 31B, 26B sizes) supporting text, audio, image inputs with up to 256K context. Available under Apache 2.0 license for local deployment on smartphones and PCs. Reported in Indian business media with confirmation from Demis Hassabis. It matters because it significantly broadens access to frontier-level open AI capabilities optimized for on-device and edge use, accelerating developer adoption and reducing reliance on proprietary cloud models.
At a Glance
Date
April 2, 2026
Importance
High
4/5
Category
model
Axis of Change
Deployment Expansion
Organizations
Models Affected
Sources