Founder of DeepSeek
Appears in 2 stories
Runs DeepSeek from Hangzhou; rarely speaks publicly
DeepSeek released V4.1-Flash on September 10, a model that reads a million tokens of context while using a fraction of the memory its predecessors needed. The key-value cache, the memory a model keeps to recall what it has already read, drops to 890 bytes per token, one quarter of the prior Flash generation.
Updated Yesterday
Leading DeepSeek's push to monetize its flagship models
DeepSeek's flagship model left preview on August 13. Independent benchmark results are mixed: V4-Pro scored 53 on the Artificial Analysis Intelligence Index, trailing a mid-tier OpenAI GPT-5.6 model but roughly matching Zhipu AI's GLM-5.2, while beating Anthropic's Claude Fable 5 on cybersecurity tests.
Updated Aug 14
No stories match your search
Try a different keyword
How would you like to describe your experience with the app today?
