According to 1M AI News monitoring, MiniMax has announced a comprehensive upgrade of its original Coding Plan to a Token Plan. Building on the M2.7 programming model, it has added support for invoking the Hailuo video model, Speech speech model, Music music model, and Image image generation model. Users can cover all modals from code writing to video generation with a single Token Plan Key.
Users on the Plus plan and above will receive a bonus quota for the multimodal model, which will not consume the original programming model usage. Usage within the plan's bonus quota does not incur additional charges. The usage and user experience of the programming model remain consistent with the original Coding Plan.
To address the computational pressure caused by the rapid increase in usage after the M2.7 launch, MiniMax has simultaneously introduced traffic control measures: dynamic throttling during peak hours from 15:00 to 17:30 on workdays and a weekly usage cap (10 times the "5-hour usage"). Developers and enterprise users with highly concurrent or batch task requirements can choose the "pay-as-you-go" mode to obtain a general API Key, which is not subject to dynamic throttling.
In addition, MiniMax has launched Voice and Video Resource Packs supporting the flagship Speech model Speech 2.8 and the video model Hailuo 2.3/2.3-Fast, which are reportedly priced to save up to 20% compared to individual calls.
