[On the First Anniversary of DeepSeek-R1 Release, New Model 'MODEL1' Unveiled] BlockBeats News, January 21: According to a report by Quantum Bit, on the first anniversary of the release of DeepSeek-R1, the new model 'MODEL1' was unveiled. DeepSeek updated the FlashMLA code on GitHub, where 'MODEL1' was mentioned 28 times across 114 files, appearing as a different 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 reflected in KV cache layout, sparsity handling, and FP8 decoding, with multiple variations in memory optimization.