Deepseek VL-2 is a scalable vision-language model using a mixture of experts (MoE) architecture to optimize performance and resource usage by activating only relevant sub-networks for specific tasks.

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

The company recently released an upgraded version of V3, a general-purpose model, and is expected to update its R1 “reasoning” model soon. Topics AI, deepseek, In Brief October 27-29, 2025 ...

Samba-1 is based on what the company describes as a composition of experts architecture. It comprises more than a half dozen open-source neural networks from OpenAI, Microsoft Corp., Meta ...

I am running Mistral 8x7B instruct at 27 tokens per second, completely locally thanks to @LMStudioAI. A model that scores better than GPT-3.5, locally. Imagine where we will be 1 year from now." ...

Chinese AI startup DeepSeek, known for challenging leading AI vendors with its innovative open-source technologies, today released a new ultra-large model: DeepSeek-V3. Available via Hugging Face ...

Touted as the “most open enterprise-grade LLM” in the market, Arctic taps a unique mixture of expert (MoE) architecture to top benchmarks for enterprise tasks while being efficient at the same ...