Show notes
Gemma 4 is not just another open model launch.
Google is packaging:
• reasoning + multi-step planning
• function calling + JSON output + system instructions
• image/video understanding across the full family
• native audio input on E2B / E4B
• 128K to 256K context
• Apache 2.0 licensing
Google says its 31B Dense is the #3 open model and 26B MoE is #6 on Arena AI as of April 1, 2026.
The bigger story: AI Edge Gallery + LiteRT-LM turn Gemma 4 into a serious on-device agent stack.
Source: https://blog.google/innovation-and-ai/technology/developers-tools/gemma-4/ Edge…
Website: podcasters.spotify.com

