Announcement
DeepSeek launches V4.1-Flash-1017 across its website, app and API
The company says the model improves benchmarks, task speed and output length over its predecessor.
TLDR
DeepSeek says V4.1-Flash-1017 is live on its website, app and API, with stronger benchmark performance, faster task completion and shorter outputs than the previous version. The Decoder reports it is open-weight, while posts sharing the announcement highlight a 77% DeepSWE 1.1 score, low off-peak pricing and a sparse 552-billion-parameter architecture. The Decoder also reports DeepSeek is opening Search and Fetch APIs to developers.
Combined views
7.5M
33 Sources, first seen ago
