Model overview
What this model is for
Google: Gemma 3 27B is a multimodal, open-weight model suited to applications that need image understanding alongside text generation. It accepts text and image inputs and produces text outputs, making it useful for visual question answering, image-based analysis, document or screenshot interpretation, and conversational interfaces that combine visual and textual context. The model record also lists tool-calling, which can support integrations that invoke external functions or services, although the supplied material does not specify supported tool schemas, reliability, or implementation details. Its advertised context window is up to 131,072 tokens, providing substantial room for long prompts and multimodal workflows; this is an advertised limit rather than a guaranteed runtime configuration. The supplied summary describes improved math, reasoning, and chat capabilities and support for more than 140 languages, but no independent benchmark results or inference-speed measurement are available in the supplied material. Listed pricing is $0.08 per million input tokens and $0.35 per million output tokens in the model record. Licensing terms, usage restrictions, safety characteristics, deployment requirements, and hardware recommendations are not available in the supplied material, so those should be verified before production use.