May 26, 2026
99% off and the comments lost it
Xiaomi MiMo-v2.5 Series API Permanent Price Reduction Up to 99%
Xiaomi slashes AI prices so hard commenters are acting like Silicon Valley just got dumped
TLDR: Xiaomi permanently cut prices on its AI service by up to 99% and boosted plan usage, with current users also getting their quotas reset. Commenters treated it like a tech industry cage match, joking that Google is doomed and arguing Chinese firms are winning by making AI cheap for everyone.
Xiaomi didn’t just trim prices on its MiMo-v2.5 AI tools — it obliterated them, announcing permanent cuts of up to 99% and sweetening the deal by resetting current user quotas and making paid plans go much further than before. In plain English: people who were already getting a pile of usage just woke up to find that pile turned into a mountain. One commenter, passive, was stunned that a plan he barely dented suddenly ballooned to 38 billion tokens a month, joking that he might actually need to downgrade because the new allowance is absurdly huge.
But the real spectacle is in the comments, where the reaction ranges from awe to full-blown geopolitical chest-thumping. One dramatic post declared, “Google was just killed,” while another framed the move as China winning the race to make AI so cheap that the expensive players can’t keep up. That sparked the spiciest mood in the thread: not just excitement over bargain prices, but the idea that Chinese AI companies are humiliating their US rivals by cutting costs while others seem to keep charging premium rates.
There’s also meme energy everywhere. People are laughing at the sheer scale of the discount — 99% sounds less like a sale and more like a dare — and joking that nobody should tell Anthropic CEO Dario Amodei. Underneath the chaos, though, the mood is clear: developers love cheap tools, and Xiaomi just turned a pricing update into a comment-section victory lap.
Key Points
- •Xiaomi said its MiMo-V2.5 Series API will receive a permanent price reduction of up to 99%, effective globally at 0:00 on May 27, Beijing time.
- •The updated Token Plan billing system increases usable volume to 5-8 times the previous level without increasing price and applies clearer billing rules.
- •Xiaomi said the Quadrillion Token Creator Incentive Program launched on April 28 and ended early after all 100T tokens were distributed by 16:08 on May 26, Beijing time.
- •All active Token Plan users, including eligible incentive-program participants and Apache Software Foundation benefit recipients, will have their Credits quota fully reset on May 27 under the new billing rules.
- •Xiaomi attributed the lower pricing to inference optimizations including SWA based on SGLang HiCache, reduced KV Cache transfer, expanded cacheable tokens, and throughput improvements from expert parallelism and input-length bucketing.