China's large models have once again unveiled a "killer feature" in their commercialization path. On March 23, MiniMax officially announced that its original coding subscription plan, Coding Plan, has been fully upgraded to the world's first multi-modal model support subscription plan — Token Plan.
The core of this upgrade lies in "comprehensiveness" and "cost reduction":
Multi-modal coverage:Token Plan is no longer limited to programming, but comprehensively integrates MiniMax's star model matrix, including Hailuo video model, Speech voice model, Music music model, and Image image generation model.
More features without additional cost: For users with Plus and above plans, the system will additionally grant multi-modal model usage quotas on top of the existing programming model quotas. Using video, audio, and image generation features within the granted quota does not require additional payment.
Latest model launch: This plan now also supports calling MiniMax's latest released M2.7 coding model.
Deep integration: OpenClaw efficiency takes off again
For power users, the most exciting aspect is the打通 of the ecosystem. Based on Token Plan, users can directly install MiniMax multi-modal toolset Skill in OpenClaw (Lobster). This means developers can now more easily call advanced functions such as custom voice tones and video creation within intelligent agents.
Flexible expansion for professional users
For professional developers and enterprises with large-scale production needs, MiniMax has also launched specialized "voice resource packages" and "video resource packages." These packages support the flagship Speech2.8 voice model and Hailuo2.3 video model. Compared to individual calls, purchasing resource packages offers about a 20% discount.
From focusing on programming to covering audio, video, and images, MiniMax is building a comprehensive AI consumption loop through Token Plan
