7 Featured Models
Reasoning-first models built for agents. Now available on web, app, and API — with state-of-the-art performance on complex multi-step tasks.
Visit →A powerful, efficient, and ultra-fast foundation language model that particularly excels in reasoning, coding, and agentic scenarios.
Visit →A trillion-parameter MoE model (32B active parameters) that integrates vision natively from pretraining — trained on ~15 trillion mixed vision and text tokens with a constant vision-text mixing ratio throughout.
Visit →The strongest model in the 30B class — offers a new option for lightweight deployment that balances performance and efficiency for resource-constrained environments.
Visit →An open-weight model that fits into a single H100 GPU — 117B total parameters with only 5.1B active parameters per forward pass, enabling efficient large-scale inference.
Visit →Packs 235B total parameters (22B active per token across 128 experts) and delivers state-of-the-art performance on instruction following, coding, math, and science benchmarks.
HuggingFace →Meta's Llama 4 series introduces natively multimodal models capable of processing both text and images from the ground up — not bolted on as an afterthought.
Visit →