DeepSeek has introduced a new AI model named 'MODEL1' on the first anniversary of its predecessor, DeepSeek-R1. The announcement was made following updates to the FlashMLA code on GitHub, where 'MODEL1' was referenced 28 times across 114 files, indicating its distinction from the existing V32 model, known as DeepSeek-V3.2. The new model features significant advancements, including changes in the key-value cache layout, improved sparsity handling, and FP8 decoding, alongside various memory optimization techniques.
DeepSeek Unveils New AI Model 'MODEL1' with Enhanced Features
Disclaimer: The content provided on Phemex News is for informational purposes only. We do not guarantee the quality, accuracy, or completeness of the information sourced from third-party articles. The content on this page does not constitute financial or investment advice. We strongly encourage you to conduct you own research and consult with a qualified financial advisor before making any investment decisions.
