Skip to content
AI Model Radar

Gemma 4 26B A4B IT

gemma-4-26b-a4b-it
Google · chat · api · seen 13m ago
GA Alert me on changes
Context
262.1K
Max output
32.8K
Input $/1M
$0.06
Output $/1M
$0.33
Modalities
text
Released
03 Apr 2026
AI summary
● machine-written

Google releases Gemma 4 26B A4B instruction-tuned model

Gemma 4 26B A4B IT is an instruction-tuned Mixture-of-Experts model from Google DeepMind with 25.2B total parameters, of which only 3.8B activate per token during inference. The model supports a 262K token context window, native function calling, and structured output. It is available via API at $0.06 per million input tokens and $0.33 per million output tokens.

What's new
  • Mixture-of-Experts architecture with 3.8B active parameters per token
  • 262K token context window with 32.8K max output
  • Native function calling and configurable thinking/reasoning mode
  • Released April 3, 2026 under Apache 2.0 license
Best for
Cost-efficient instruction following at scaleLong-context document processingAPI-based chat and instruction completion
Sources

Source: https://ai.google.dev/gemini-api/docs/models/gemma-4-26b-a4b-it