← Back to all models
💬

Google: Gemma 4 26B A4B (free)

Google·Text Generation
🔥 50trending

Gemma 4 26B A4B IT is an instruction-tuned Mixture-of-Experts (MoE) model from Google DeepMind. Despite 25.2B total parameters, only 3.8B activate per token during inference — delivering near-31B quality at...

#text+image+video->text#top-provider
🧮

Undisclosed

Parameters

📏

262K tokens

Context Window

🔒

Proprietary

License

📅

Apr 3, 2026

Released

🔌

API Available

This model is accessible via API for integration into your applications.