Xiaomi's MiMo platform has announced sweeping price reductions for its V2.5 AI model, cutting costs by up to 99% while simultaneously expanding token plan quotas by factors of five to eight times. The changes take effect May 27, 2026 at 00:00 AM China Standard Time, marking one of the most aggressive pricing maneuvers in the competitive AI inference space this year.

Context: An AI Pricing Arms Race

The announcement arrives amid intensifying competition among AI infrastructure providers, where major players have been steadily reducing per-token costs to attract developers and enterprise customers. Xiaomi's dramatic cut positions MiMo as a budget-conscious alternative to pricier rivals, potentially reshaping how smaller operators compete in the crowded LLM-as-a-service market.

What's Changing for MiMo-V2.5 Users

Beyond the headline 99% price reduction, existing Token Plan subscribers will see their credit allocations expanded significantly—between five and eight times current limits. Perhaps most notably, all previously used credits within the validity period will be fully reset, giving users a clean slate when the new pricing structure activates.

Developer Implications

For builders working with AI APIs, such drastic cost reductions could fundamentally alter project economics. Applications previously deemed too expensive to scale may become viable, while existing MiMo users stand to benefit from immediate savings on ongoing workloads. The reset provision is particularly significant for those who burned through their quotas earlier in the billing cycle.

What Remains Unclear

The source announcement lacks granular details—specific per-1K-token pricing tiers, new credit expiration windows, and whether the cuts apply uniformly across all model variants remain unknown. Developers interested in the changes should monitor Xiaomi's official documentation for full pricing breakdowns before migrating workloads.

Key Takeaways

  • MiMo-V2.5 prices drop up to 99% starting May 27, 2026 at midnight CST
  • Token Plan quotas increase by 5–8× with a full credit reset on activation
  • The aggressive pricing signals Xiaomi's push for market share in AI inference
  • Specific per-token costs and model variant details await official documentation

The Bottom Line

Xiaomi's aggressive price cuts signal that the AI inference market is entering a commodity phase where cost efficiency matters more than raw capability. Developers should treat this as a signal to audit their API spending—but wait for official pricing docs before committing workloads.