BlockBeats News: On January 21, according to QbitAI, on the one-year anniversary of the release of DeepSeek-R1, the company revealed a new model called "MODEL1." DeepSeek updated the FlashMLA code on GitHub, and across 114 files, "MODEL1" was mentioned 28 times, appearing as a distinct model from V32. It is known that V32 refers to DeepSeek-V3.2, and MODEL1 is likely a new architecture. Specific differences in the code are evident in the key-value (KV) cache layout, sparsity handling, and FP8 decoding, with multiple variations in memory optimization.
DeepSeek Reveals New Model 'MODEL1' One Year After the Launch of DeepSeek-R1
KuCoinFlashShare
Summary
On January 21, 2026, DeepSeek announced a new model called "MODEL1," exactly one year after the launch of DeepSeek-R1. The company updated the FlashMLA code on GitHub, where MODEL1 was mentioned 28 times across 114 files, clearly distinguishing it from V32 (DeepSeek-V3.2). Code changes indicate differences in KV cache layout, sparsity handling, and FP8 decoding, suggesting improvements in memory optimization. This update comes amid ongoing news about token launches and new token listings on major platforms.
Source:Show original
Disclaimer: The information on this page may have been obtained from third parties and does not necessarily reflect the views or opinions of KuCoin. This content is provided for general informational purposes only, without any representation or warranty of any kind, nor shall it be construed as financial or investment advice. KuCoin shall not be liable for any errors or omissions, or for any outcomes resulting from the use of this information.
Investments in digital assets can be risky. Please carefully evaluate the risks of a product and your risk tolerance based on your own financial circumstances. For more information, please refer to our Terms of Use and Risk Disclosure.