Skip to main content

MiniMax Backtracks on Pricing Change After User Outcry

MiniMax Stumbles in Pricing Transition, Offers Make-Good Plan

The AI industry got another lesson in customer communication this week as MiniMax faced a developer revolt over unexpected changes to its pricing model. The company's quiet shift from task-based to token-based billing for its new MiniMax-M3 model left many users feeling blindsided.

What Went Wrong?

When MiniMax launched its flagship M3 model last week, it replaced the familiar Coding Plan with a new Token Plan system. The change came with little warning and hit heavy users particularly hard - many burned through their monthly allocations in days rather than weeks. Developer forums quickly filled with complaints about the opaque transition and unexpectedly high token consumption.

"We recognize we mishandled this transition," the company admitted in a June 2 statement. MiniMax acknowledged failures in user communication, transition planning, and treatment of existing customers' usage limits.

Why the Change?

The new M3 model represents a significant technical leap, supporting native multimodal capabilities and an impressive 1 million token context window. These advancements come with substantially higher computational demands. MiniMax argues that token-based billing better accommodates users who want flexibility across different modalities.

Making Things Right

Facing mounting criticism, MiniMax rolled out a four-part compensation plan:

  • Legacy Benefits Protected: Early adopters (pre-March 2026) keep their unlimited weekly access for both M2.7 and M3 models
  • Permanent Limit Boost: Token Plan subscribers get a 50% weekly limit increase for M3
  • Temporary Bonus: All users receive double quota (10 hours/week) through June 7
  • Extended Redemption: Compensation points now valid for a full year, with self-service refunds available

Broader Industry Trend

This isn't the first pricing controversy in the AI space. Moonshot (Kimi) faced similar protests recently. As models evolve into AI agents capable of complex, multi-step tasks, token consumption has skyrocketed. While token-based billing may be economically necessary, companies are learning the hard way that transparency matters just as much as technology.

Key Points

  • MiniMax changed billing from per-task to token-based without proper user communication
  • Heavy users exhausted quotas faster than expected, sparking backlash
  • Company admits missteps, offers compensation package including:
    • Preserved benefits for early adopters
    • Permanent 50% limit increase for some subscribers
    • Temporary quota doubling through June 7
    • Extended point validity and refund options
  • Incident highlights growing pains as AI models become more resource-intensive