MiniMax Launches Token Plan: World's First Subscription for Full-Modal AI Models
Category: Tool Dynamics
Excerpt:
Chinese AI company MiniMax launches Token Plan subscription plan, becoming the world's first unified subscription service that supports full modal models. This plan covers five major modes: text, voice, video, image, and music. The monthly fee starts at $10 and provides M2.7-high high-speed inference services with a throughput of 100 TPS.
Shanghai — March 23, 2026 — Chinese AI company MiniMax today announced Token Plan, the world's first unified subscription plan supporting full-modal AI models. The new offering consolidates access to text, speech, video, image, and music generation under a single monthly subscription starting at $10, marking a significant shift in how developers access and pay for multimodal AI capabilities.
📌 Key Highlights at a Glance
- Product: Token Plan — World's first unified subscription for full-modal AI
- Developer: MiniMax (Hong Kong listed: 00100.HK)
- Modalities: Text, Speech, Video, Image, Music
- Starting Price: $10/month
- High-Speed Plans: Plus, Max, Ultra tiers at $40/month
- Performance: M2.7-highspeed delivers ~100 TPS sustained throughput
- Speed Advantage: Up to 3× faster than competing models
- Replacement: Upgrades previous Coding Plan to full multimodal access
- Availability: Immediate access via MiniMax API Platform
🎫 What is Token Plan
Token Plan represents a paradigm shift in AI service pricing. Unlike traditional pay-per-token models or single-modality subscriptions, Token Plan provides unified access to MiniMax's complete model portfolio covering all five major AI modalities—text, speech, video, image, and music—under a single monthly fee.
The announcement marks an upgrade from MiniMax's previous Coding Plan, expanding coverage from code-focused models to the company's entire multimodal ecosystem. For developers, this consolidation eliminates the complexity of managing multiple subscriptions across different AI services while significantly reducing costs.
Token Plan Core Features
Full-Modal Access
Single subscription covers text, speech, video, image, and music models
Fixed Monthly Fee
Predictable pricing starting at $10/month with no usage-based surprises
High-Speed Options
Dedicated M2.7-highspeed support for faster inference needs
Auto-Reset Quotas
M2.7 requests reset every 5 hours for continuous access
"Token Plan supports MiniMax models across all modalities—text, speech, video, image, and music. A fixed-fee subscription grants you access to the entire model matrix."
— MiniMax API Documentation, March 2026
💵 Pricing Tiers and Options
Token Plan offers flexible tiers designed to accommodate different usage levels and performance requirements:
Token Plan Pricing Structure
| Plan Type | Tier | Monthly Price | Key Features |
|---|---|---|---|
| Standard | Starter | $10 | Basic multimodal access, standard speed |
| Plus | $20 | Higher quotas, priority access | |
| Max | $30 | Maximum standard-tier quotas | |
| High-Speed | Plus-Highspeed | $40 | M2.7-highspeed, ~100 TPS throughput |
| Max-Highspeed | $40+ | Higher quotas with high-speed inference | |
| Ultra-Highspeed | $40+ | Maximum performance tier |
Cost Efficiency Analysis
Compared to pay-per-token pricing, Token Plan delivers significant savings for developers with consistent usage patterns. MiniMax's M2.5 model already offers API pricing at one-tenth to one-twentieth the cost of competitors like Claude Opus, Gemini 3 Pro, and GPT-5—Token Plan extends this value proposition with predictable monthly fees.
- M2.7 API Pricing: $0.30/million input tokens, $1.20/million output tokens
- Token Plan Value: Fixed monthly cost eliminates variable pricing uncertainty
- High-Speed Premium: $40/month for 3× faster inference
🎨 Full-Modal AI Capabilities
MiniMax's model matrix covers five distinct modalities, enabling developers to build comprehensive AI applications through a single API:
📝 Text
Large language models with ultra-long context processing (up to 4 million tokens), advanced coding capability, and high agentic performance
🎤 Speech
Text-to-speech and speech-to-text models supporting multiple languages and natural voice synthesis
🎬 Video
Video generation and understanding models for content creation and analysis applications
🖼️ Image
Image generation, editing, and understanding capabilities integrated with text models
🎵 Music
AI music generation for creative applications and multimedia content production
Cross-Modal Integration
The unified subscription enables seamless cross-modal workflows. Developers can build applications that combine text understanding with image generation, speech synthesis with video creation, or any combination of modalities—all within a single subscription framework. This consolidation simplifies architecture, reduces integration complexity, and enables more sophisticated AI applications.
⚡ Performance and Speed
Token Plan's High-Speed tiers leverage MiniMax's latest M2.7-highspeed model, delivering performance that significantly exceeds competing offerings:
M2.7-Highspeed Performance
| Metric | Value | Comparison |
|---|---|---|
| Sustained Throughput | ~100 TPS | Industry-leading for subscription models |
| Speed vs Competition | Up to 3× faster | Compared to competing models at similar price points |
| Context Window | 197K tokens (M2.5) | Ultra-long context processing |
| Request Reset | Every 5 hours | M2.7 requests auto-refresh |
Speed Tier Benefits
- Dedicated Infrastructure: High-speed subscriptions allocate dedicated M2.7-highspeed resources
- Consistent Latency: Predictable response times for production applications
- Higher Throughput: 20× the throughput of Starter tier for demanding workloads
- Priority Queue: Requests processed with priority during peak usage
📊 Market Comparison
Token Plan enters a market dominated by single-modality subscriptions and pay-per-token pricing. MiniMax's unified approach addresses several key pain points for developers:
Subscription Model Comparison
| Provider | Modalities | Model | Starting Price |
|---|---|---|---|
| MiniMax Token Plan | Text, Speech, Video, Image, Music | M2.7, M2.5, Full Matrix | $10/month |
| OpenAI | Text, Image, Audio | GPT-5.4, DALL-E, TTS | Pay-per-use |
| Anthropic | Text | Claude 4.6 | Pay-per-use |
| Text, Image, Video | Gemini 3, Veo 3, Imagen | Pay-per-use |
Market Impact
- Pricing Pressure: Fixed-fee subscription challenges pay-per-token industry standard
- Consolidation Trend: Unified subscriptions reduce vendor management complexity
- Developer Experience: Single API for all modalities simplifies development
- Cost Predictability: Monthly fees enable better budget planning for businesses
👨💻 Developer Access
Token Plan is immediately available through the MiniMax API Platform:
Getting Started
Create account at platform.minimax.io
Select tier at platform.minimax.io/subscribe/token-plan
Obtain credentials for API authentication
Access text, speech, video, image, and music through unified API
❓ Frequently Asked Questions
What is MiniMax Token Plan?
Token Plan is MiniMax's unified subscription service providing access to all MiniMax AI models across five modalities—text, speech, video, image, and music—under a single monthly fee starting at $10. It replaces the previous Coding Plan with expanded multimodal coverage.
Which models does Token Plan support?
Token Plan supports all MiniMax models across all modalities: text (M2.7, M2.5), speech (TTS, STT), video generation, image generation/editing, and music generation. M2.7 requests reset every 5 hours for continuous access.
How much does Token Plan cost?
Token Plan starts at $10/month for the Starter tier. Standard plans range from $10-30/month, while High-Speed plans (M2.7-highspeed with ~100 TPS throughput) start at $40/month. All plans provide access to the full model matrix.
What is the difference between Standard and High-Speed plans?
High-Speed plans provide dedicated M2.7-highspeed model access with approximately 100 TPS sustained throughput—up to 3× faster than competing models and 20× the throughput of the Starter tier. Standard plans use regular inference speeds suitable for most applications.
How does Token Plan compare to pay-per-token pricing?
Token Plan offers predictable monthly costs instead of variable usage-based pricing. For developers with consistent usage patterns, this typically provides better value. MiniMax's models already offer significantly lower pricing than competitors—Token Plan extends this value with fixed-fee predictability.
🎤 Industry Perspectives
"Token Plan represents a significant shift in AI pricing models. Unified access to all modalities under a single subscription eliminates the complexity developers face when integrating multiple AI services."
— AI Industry Analyst, March 2026"MiniMax has been aggressive on pricing with M2.5 at one-tenth the cost of competitors. Token Plan extends this strategy with predictable monthly fees that make budgeting straightforward for businesses."
— Technology Review, March 2026"The world's first full-modal subscription is a notable milestone. Whether competitors follow with similar unified offerings will be interesting to watch."
— Developer Community Voice, March 2026The Bottom Line
MiniMax's Token Plan represents a significant innovation in AI service pricing. By consolidating access to all five modalities—text, speech, video, image, and music—under a single monthly subscription starting at $10, MiniMax has created what may become a new industry standard for AI service delivery.
For developers, the benefits are clear: simplified vendor management, predictable costs, and unified API access across all modalities. The High-Speed tiers with M2.7-highspeed's ~100 TPS throughput address the needs of production applications requiring consistent performance.
For the broader market, Token Plan puts pressure on competitors still relying on complex pay-per-token pricing across fragmented service offerings. As AI becomes infrastructure for more applications, the simplicity of unified subscriptions may prove increasingly attractive.
MiniMax, already recognized as one of China's "AI Four Little Dragons" and publicly traded on the Hong Kong Stock Exchange (00100.HK), continues its strategy of aggressive pricing and developer-friendly offerings. Token Plan is another step in making advanced AI capabilities accessible to a broader developer community.
Stay tuned to our Tool Dynamics section for continued coverage of AI pricing and platform developments.
