Data compiled by Woofun AI shows that MiniMax has enforced a permanent 50% price reduction on its flagship MiniMax-M3 model, just two weeks after its initial deployment. Under the standard pay-as-you-go structure, the cost for one million input tokens within the 512k context window has dropped from $0.60 to $0.30, while output pricing fell from $2.40 to $1.20. This adjustment also applies to inputs exceeding 512k tokens, effectively halving rates for extra-long context processing.

This strategic pivot follows a significant trust crisis initiated on June 1st, when MiniMax abruptly transitioned from a pay-per-call mechanism to token-based billing, driving existing user costs up by over 250%. Although parent company RareSky Technology issued an apology on June 2nd, resetting quotas and offering compensation, developer exodus persisted due to intense competition from rivals like DeepSeek. Consequently, the permanent price cut was implemented on June 15th.

The aggressive pricing strategy has triggered volatility in the secondary capital market. Goldman Sachs warned that the move would severely erode profit margins, resulting in a 14% reduction of its target price for MiniMax.

Concurrently, JPMorgan lowered its rating, arguing that rapid post-launch price cuts often indicate that a model's actual capabilities fall short of expectations. Investors have responded by significantly favoring short positions on MiniMax in the Hong Kong stock market.