Compared to DeepSeek R1, Llama-3.1-Nemotron-Ultra-253B shows competitive results despite having less than half the parameters.

The Llama 4 series is the first to use a “mixture of experts (MoE) architecture,” where only a few parts of the neural ...

Meta has launched Llama 4 Scout, Maverick, and Behemoth models, offering open source, multimodal AI with long context and ...

Research from Stack Overflow indicates a strong overall preference from developers for open-source AI models and tools.

"The vibes around llama 4 so far are decidedly mid ," independent AI researcher Simon Willison told Ars Technica. Willison ...

Meta and OpenAI's adoption of open-weight models signals a shift from the dominance of closed-source AI models in the West.

The new family of models comes as Meta works to strip bias from its AI and ensure it can engage with controversial topics.

Consisting of Llama 4 Scout, Llama 4 Maverick, and Llama 4 Behemoth, each designed with different capabilities and scales.

Meta has unveiled Llama 4, its latest family of large language models (LLMs). These models now power the Meta AI assistant across platforms such as ...

The folks over at buzzy Chinese AI lab, DeepSeek, are working on a new series of AI models called DeepSeek-GRM that employ a ...

Meta has also previewed Llama 4 Behemoth, the largest AI model in the family so far, with 288 billion active parameters.