DeepSeek released the DeepSeek-V4.1-Flash-1017 model, featuring stronger benchmark performance, faster task completion, and shorter outputs, described as a Pareto improvement in performance and cost. It is now live on the official website, app, and API, with optimizations for DeepSeek Harness, Tencent CodeBuddy, WorkBuddy, and more.
The X account AI Tracker Bot spotted an entry named "deepseek-v4.1-flash-1017" in DeepSeek's official API list before it was removed. The account speculated the 1017 suffix points to a new V4.1 Flash snapshot, but the public deepseek-flash remains V4.1 Flash, and DeepSeek has not yet commented.
DeepSeek V4.1 Pro has reportedly entered early access, rolling out gradually by account with a default High reasoning intensity. Users who gain access can continue using it, and the company has not yet announced a definitive release date. The model was previously mentioned by DeepSeek when it released V4.1 Flash on September 10.
OpenCode launched the second phase of "Operation Cheepseek" on September 25, making the $60 credit for DeepSeek V4.1 Flash on OpenCode Go permanent. The credit had been temporarily raised from $15 to $60 and was scheduled to end on September 20; the Go page now shows $60 with the original $15 crossed out.
Deep Code, an open-source terminal programming assistant, gains attention for deep integration with DeepSeek-V4 models, excelling in reasoning and code generati
The price of the DeepSeek-V4-Pro model API will be officially adjusted to 1/4 of the original price after the 25% discount ends on May 31, 2026.
Tencent Cloud announced that its two core large models, Hy3preview and DeepSeek-V4-Pro, on its agent development platform will end free public beta testing on M