SambaNova Systems
September 17, 2024
SambaNova Systems Unicorn News - September 17, 2024
SambaNova Systems has introduced a high-speed demo powered by Meta’s Llama 3.1 Instruct model, challenging OpenAI’s o1 model. This new offering aims to provide developers with a fast, open-source alternative, positioning SambaNova as a key player in the generative AI market.
() Introduction
() SambaNova Unveils High-Speed Llama 3.1-Powered Demo, Challenging OpenAI's O1 Model
SambaNova Systems has launched a new demo on Hugging Face, utilizing Meta’s Llama 3.1 Instruct model. Announced on September 16, 2024, this demo offers a high-speed, open-source alternative to OpenAI’s recently released o1 model. It allows developers to interact with the Llama 3.1 405B model at speeds of 129 tokens per second. Initial benchmarks reveal that the demo, running on SambaNova’s infrastructure, achieved 405 tokens per second for the Llama 3.1 70B model, making it the second-fastest provider of Llama models, just behind Cerebras. This development highlights SambaNova’s strategy to expand its presence in the generative AI market by providing a highly efficient and scalable platform for both developers and enterprises.
