"DeepSeek's announcement heralds the arrival of cost-effective AI models with a context length of up to 1 million tokens, enhancing coherence in extended conversations."
"The V4 Pro model is said to rival leading closed-source models in reasoning capabilities, while V4 Flash offers quicker responses and comparable performance on simpler tasks."
"Following its viral success, DeepSeek faced bans from US federal agencies due to perceived national security risks, and South Korea paused app downloads over privacy issues."
DeepSeek has introduced its V4 Pro and Flash AI models, boasting a context length of up to 1 million tokens for improved conversation coherence. The V4 Pro model offers enhanced reasoning capabilities, rivaling top closed-source models, while V4 Flash provides faster response times with similar reasoning abilities for simpler tasks. Despite its success, DeepSeek faced bans from US federal agencies and South Korea due to national security and privacy concerns, following its rise to the top of the App Store charts.
Read at Engadget
Unable to calculate read time
Collection
[
|
...
]