Gemma 4 26B A4B

Gemma 4 26B A4B IT is an instruction-tuned Mixture-of-Experts (MoE) model developed by Google DeepMind. While it has 25.2B parameters in total, only 3.8B are activated per token during inference — providing performance close to a 31B model at a much lower computational cost. It supports multimodal inputs such as text, images, and video (up to 60 seconds at 1 fps). Key features include a 256K-token context window, native function calling, an adjustable thinking/reasoning mode, and support for structured outputs. It is released under the Apache 2.0 license.