Skip to content
AI Model Radar

Gemma 4 31B IT

gemma-4-31b-it
Google · chat · api · seen 11m ago
GA Alert me on changes
Context
262.1K
Max output
32.8K
Input $/1M
$0.12
Output $/1M
$0.35
Modalities
text
Released
02 Apr 2026
AI summary
● machine-written

Google releases Gemma 4 31B Instruct, 262K-token chat model

Gemma 4 31B Instruct is Google DeepMind's 30.7B parameter instruction-tuned model supporting text input and output. Released April 2, 2026, it offers a 262K token context window and is available via API at $0.12 per million input tokens and $0.35 per million output tokens through Google.

What's new
  • 262,144 token context window
  • Supports vision input and tool calling capabilities
  • Output speed of 35 tokens/second with 0.96s median time-to-first-token
  • Available through multiple providers including Google, DeepInfra, and OpenRouter
Best for
Long-context text generation tasksVision-enabled applications requiring image and text inputTool-calling and structured output (JSON schema) workflowsCost-sensitive applications with moderate reasoning requirements
Sources

Source: https://ai.google.dev/gemini-api/docs/models/gemma-4-31b-it