DeepSeek released DeepSeek-V4.1-Flash-1017 on October 17. Two tools are now letting people use DeepSeek V4.1 Flash without paying, but the terms are different: one gives a date, the other gives a daily allowance.
WorkBuddy Intl: free through October 31
WorkBuddy posted on X on October 17 that "WorkBuddy Intl is updated to the latest too, and it's free through Oct 31." The post quoted DeepSeek's announcement of DeepSeek-V4.1-Flash-1017. The post refers only to the international version and doesn't give any usage limits.
According to a third-party summary on AI Promo Now, a notice in the WorkBuddy client says the model is paused for an account after about 200 million tokens, cache included, and the free allowance resets automatically at the time given in the notice. WorkBuddy hasn't published this on its own pages.
Huawei Cloud CodeArts: 10 million tokens a day
Huawei Cloud's help center says all users of its CodeArts coding agent can claim 10 million tokens a day. The tokens are cleared at 24:00 Beijing time the same day and don't carry over. The event's models are deepseek-v4.1-flash, glm-5.3-flash, deepseek-v4-flash-0731 and deepseek-v4-pro-0813. The page doesn't give a version number for deepseek-v4.1-flash, so it doesn't say whether it's the 1017 release.
The page lists the end date as "until further notice," not October 31. The tokens are claimed automatically after you log in to the CodeArts IDE, the JetBrains or VS Code plugin, the desktop app, the web app or the mini program, and then you switch to one of the event models. The page says these models only use the free tokens and can't draw on credit packs, so once a day's tokens run out they can't be used until the next day's grant. Huawei's CodeArts Agent event page adds that the tokens can only be used inside CodeArts and are first come, first served.
Search and fetch tools in Claude Code and Codex
DeepSeek's docs say server-side search and fetch calls made through DeepSeek Harness, Claude Code or Codex are free until 00:00 Beijing time on November 1, when per-request billing starts. Another of our articles covers those tools in detail.