Gemma 4 31B Now Available

ChatBotKit now supports Gemma 4 31B Instruct, Google DeepMind's latest open model and a versatile addition to the platform's model lineup. The model is available immediately across bots, blueprints, skillsets, and the API.

Gemma 4 31B is a multimodal model that accepts both text and image inputs, making it practical for document analysis, visual reasoning, and workflows where users need to reference screenshots, diagrams, or photographs alongside their queries. Its native function calling support means it integrates cleanly with ChatBotKit Skillsets and Blueprint-based agent architectures without additional adaptation.

The model's 262,144-token context window is one of the larger windows available on the platform, enabling analysis of long documents, extended conversation histories, and complex codebases within a single context. Gemma 4 31B also supports configurable reasoning - useful for tasks where explicit chain-of-thought processing improves accuracy - and delivers multilingual coverage across more than 140 languages, making it a practical choice for international deployments.

At its price point, Gemma 4 31B offers strong value for teams building applications that require image understanding or broad language coverage without the cost profile of a frontier model. It is a solid general-purpose option for content moderation, multilingual customer support, visual data extraction, and conversational agents that need to handle diverse input types.

Select Gemma 4 31B from the model picker in your dashboard or reference it through the API.