ggml-org/gemma-4-12b-it-GGUF · Hugging Face
Google DeepMind has released Gemma 4, a new family of open-weight multimodal models. These models support text and image inputs, with some variants also handling audio and video. Gemma 4 models feature a large context window of up to 256K tokens and are available in various sizes and architectures, including Dense and Mixture-of-Experts, making them deployable across a range of devices. AI
IMPACT Expands access to multimodal AI capabilities, enabling new applications in text, image, and audio processing across various devices.