Image: TechcityAuthority
Moonshot AI's new Kimi K3.1 Pro has been out for less than a day, and the loudest talk about it on X is not the benchmarks. It is the usage limits. A wave of paying subscribers say the model burns through their quota far faster than they expected. The model launched on October 23.
The complaints started on launch night. One account, StatsWire, claimed people were using up a full 5-hour window in under 10 minutes. It also said thinking mode is forced on and that the faster Highspeed version doubles the burn rate.
Kimi K3.1 Pro launch night and the quota is already a disaster.
— @StatsWire October 23, 2026
People are burning their entire 5-hour window in under 10 minutes.
Some in less than 2.
The most shared complaint came from Da7_Tech, with more than 300 likes. The poster praised K3.1 Pro as the best open model available but called it practically unusable on its own subscription. On the $39 plan, the poster said, a single test build used 11.6% of the monthly quota.
Kimi K3.1 Pro is the best open model in the world right now. No contest. But it's practically unusable on its own subscription.
— @Da7_Tech October 23, 2026
Others reported similar experiences. One user said a $31 plan hit its 5-hour limit mid-session while the coding tool showed only 42.8K context tokens, and called it a possible quota bug. Another warned people not to subscribe, saying they ran into both monthly and weekly caps and saw no usage resets.
Benchmarks aside, Kimi's subscription quotas look broken.
— @thegitai October 23, 2026
A $31 plan exhausted its entire 5-hour limit mid-session, while the CLI showed just 42.8K context tokens.
For everyone considering Kimi Subscription: DO NOT SUBSCRIBE. They have Monthly and weekly limits, BOTH.
— @moreinmotion October 23, 2026
These are individual accounts. None of the numbers above have been independently checked, and a handful of angry posts cannot show how most subscribers are faring. Usage also depends heavily on the task, the thinking level and the tools people connect.
Moonshot has not publicly responded to the limit complaints on X at the time of this writing. Its official posts since the launch have focused on benchmarks and demos. Its published plan rules do explain some of what users are hitting, though. According to the Kimi Code membership page, new plans keep a rolling 5-hour rate window and drop the weekly cap. Older legacy plans keep both. Every plan also shares one monthly credit pool across Kimi on the web, Kimi Work and Kimi Code. Once that pool runs dry, coding access freezes until the monthly reset or an upgrade.
Model access is tiered as well. Moonshot's K3.1 Pro pricing page lists Plus at an effective $15 a month on annual billing, with Pro at $31, Max at $79 and Ultra at $159. Plus and above get K3.1 Pro in Kimi Code. The Highspeed version and the full 1 million token context need Pro or higher. The company's release notes also confirm that K3.1 Pro only runs in thinking mode, with high as the default level, which may explain why some users see heavy usage.
One post from Beijing-based Guanghua_David framed it this way. The poster argued that the model and the quota system are two different stories, and that the useful detail is which limit a user actually hits.
Overheard in Beijing: English X is roasting Kimi K3.1 Pro for usage limits, and barely talking about the model.
— @Guanghua_David October 24, 2026
Not everyone is down on Moonshot. Before the launch, some users praised the company for pausing new sign-ups during the K3 rush rather than cutting quotas for paying users. For now, anyone weighing a Kimi plan for heavy coding work may want to start small, watch the usage dashboard, and see whether Moonshot adjusts its limits in the coming days.
Comments