Meta Releases Llama 4.1 – Open-Source Model Tops New Reasoning Benchmarks
Meta has launched Llama 4.1, its most capable open-source model yet, which now outperforms several closed frontier models on advanced reasoning, long-context understanding, and agentic task benchmarks. The 405B-parameter model includes improved multimodal support and is being made available with full weights on Hugging Face under a permissive license, accelerating enterprise and developer adoption worldwide.
