dynamicagency/gemma-4-26b-a4b-it-oQ8e-mtp

A community-packaged variant of Google DeepMind's Gemma 4 26B A4B IT model, an open-weight multimodal Mixture-of-Experts instruction model with approximately 3.8B active parameters per token, 256K context window, and support for text and image inputs. The 'oQ8e-mtp' suffix indicates a quantized or adapted packaging of the upstream base checkpoint, hosted on Hugging Face by the dynamicagency namespace. Source: supabase.