AI NewsWaveDeepSeek Releases V4.1 Flash Update, Boosting Coding Performance and Cutting Task Time
Key points
- DeepSeek releases updated version of V4.1 Flash
- Better coding and agent performance, shorter task time
- API pricing unchanged, search API also released
Chinese AI startup DeepSeek released a new, updated version of V4.1-Flash. IT Chosun reported that the V4.1-Flash-1017 version was released on the 17th and described it as a step up in performance from the V4.1-Flash introduced last month. AI Times also reported on the update.
According to IT Chosun, DeepSeek said on its official website that benchmark performance improved over the previous version, tasks complete faster and output time is shorter. The outlet highlighted improved coding and reasoning performance in its headline.
AI Times described it as a model that cuts task time and output length while improving coding and agent performance. The outlet reported that it keeps API pricing unchanged but focuses on lowering actual usage costs by reducing the tokens needed for a task, and said in its headline that a search API was also released.
Chosun Biz reported on the 16th that DeepSeek has suffered five service outages in October alone. The outlet said DeepSeek has recently been deepening cooperation with Huawei, which makes Chinese AI chips, to reduce its reliance on Nvidia, and noted that the continuing outages have raised concerns about service stability.