Introducing Gemma 4 models on Amazon Bedrock
TL;DR
AWS is adding Google DeepMind's Gemma 4 family to Amazon Bedrock: three instruction-tuned Apache 2.0 open-weight models named Gemma 4 31B, 26B-A4B, and E2B. All three variants support text and image input, built-in reasoning mode, and native function calling. The larger 31B and 26B-A4B models offer context windows up to 256K tokens. The lineup mixes dense and MoE designs. Gemma 4 26B-A4B has 25.2B total parameters but activates 3.8B per token, while E2B targets low-cost, low-latency multimodal jobs.
Nauti's Take
Open weights get truly useful when they hit boring enterprise reality: IAM, data controls, service tiers, API compatibility. Gemma 4 on Bedrock is not hobbyist candy.
It is another lever for pushing agent stacks out of notebooks and into controlled production.
Briefingshow
This matters for teams that want open-weight models without running the serving stack, access control, and scaling layer themselves. Gemma 4 becomes less of a lab model and more of a production option inside AWS environments. The real test is whether its cost, latency, and quality beat existing Bedrock choices in actual workloads.
Sources
