Cerebras Hits 969 Tokens/Sec on Llama 3.1 405B Inference
Cerebras CS-3 just measured 969 tokens/sec on Llama 3.1 405B and ~3000 t/s on gpt-oss-120B, leaving Groq at ~476. What it means and how to use it.
Meta AI
Meta Llama, Muse Spark, Hatch, ARI, and Meta AI infrastructure.
Meta’s Hatch Personal Agent and Muse Spark Land in Beta
Meta is shipping Hatch, a consumer AI agent on the new Muse Spark model, with Instagram agentic shopping and a Q4 launch. What’s new and how to prepare.
Meta Llama 3.1: The Complete Guide to the World’s Most Popular Open-Source LLM
What Is Llama 3.1? Llama 3.1 is Meta’s (Facebook’s parent company) flagship open-source large language model. “Open-source” means anyone can