Industry-leading natively multimodal model for image and text understanding.
Meta's Llama 4 Maverick model release for language tasks, with native multimodal text and image understanding, 17B active parameters, 128 experts, and a 10M-token context window.
