[DeepSeek Updates New Model with Context Reaching Million-Level Tokens] Multiple users have reported that DeepSeek has updated its versions on both the web and app platforms, now supporting a maximum context length of 1M (one million) tokens. The context length of DeepSeekV3.1, released last August, was extended to 128K. During a test conducted by the reporter, DeepSeek claimed in its Q&A that it supports a 1M context, capable of processing ultra-long texts in one go. The reporter submitted a document of the novel *Jane Eyre* containing over 240,000 tokens, and DeepSeek successfully recognized the document's content. (Sci-Tech Innovation Board Daily)