DeepSeek V4.1 Flash new version released, wish-making programming comes true with one click. Netizens praised: Too progressive, super killing line

Kuai Technology reported on October 17th that according to the DeepSeek official public account today, DeepSeek V4.1 Flash has been updated to version V4.1-Flash-1017, and the official web page, APP and API models have been simultaneously changed to the new model. According to the official introduction, DeepSeek V4.1 Flash 1017 improves the ability to build complete projects from scratch in aspects such as game development and 3D modeling through self-feedback-based reinforcement learning (RL) training technology, and can support the realization of more "wishing-style" programming tasks and project development requests. Not only that, this new version of the DeepSeek API adds support for the Search and Fetch interfaces, and the related capabilities are officially open to the public. It is understood that the price of both interfaces is 5 yuan/thousand times. For detailed usage, please refer to the official API document: API Guide - Web Search and Crawl. The official also reminds: When current users use the DeepSeek API in DeepSeek Harness / Claude Code / Codex, they will already access the Search/Fetch interface through server-side tool calls. According to reports, this part of the Search/Fetch interface calls are currently not billed, and will be charged on a per-call basis in two weeks (00:00 on November 1, 2026, Beijing time). After the new version of DeepSeek V4.1 Flash was launched, netizens in the comment area flooded the screen with angry praises. Some netizens joked, "I just know that you were secretly making progress behind our backs when you hung up that day. You are so progressive." Some netizens said, "x x What a super kill line." Regarding the actual performance of the new version, Kuai Technology will continue to follow up and bring follow-up reports. Attached is the open source address of the model DeepSeek-V4.1-Flash-1017. The download address of the model weights is as follows: HuggingFace: https://huggingfa