gemma 4 31B it
google/gemma-4-31B-it
Open Source · chat · open-weights
Context
262.1K
Max output
8.2K
Input $/1M
$0.10
Output $/1M
$0.35
Modalities
text · image
Released
11 Mar 2026
License: apache-2.0 · google/gemma-4-31B-it
AI summary
● machine-written
Google's Gemma 4 31B Instruct multimodal model released
Gemma 4 31B Instruct is a 30.7B parameter dense multimodal model from Google DeepMind supporting text and image inputs with text output. It features a 262,144 token context window, native function calling, configurable reasoning mode, and multilingual support across 140+ languages. The model is positioned for coding, reasoning, and document understanding tasks.
What's new
- 30.7B dense multimodal architecture with text and image input support
- 262,144 token context window with configurable thinking/reasoning mode
- Native function calling and multilingual support across 140+ languages
- Apache 2.0 licensed open-weights model from Google DeepMind
- Available through 17 providers starting at $0.10 per million input tokens
Best for
Coding and technical problem-solvingReasoning and long-context understandingDocument analysis and understandingMultilingual applications