The provided document details the offerings and pricing structure for the Xiaomi MiMo artificial intelligence platform, encompassing various large language models (LLMs), multimodal capabilities, and auxiliary services.

Core Services and Functional Modules:
The MiMo platform provides a comprehensive suite of AI tools, including:

1. Chat: Core conversational AI capabilities.
2. API Integration: Facilitates seamless integration with external applications via OpenAI API, Anthropic API, and other specialized configurations (e.g., OpenCode, Claude Code, OpenClaw, Hermes Agent, Kilo Code, Cherry Studio, Qwen Code, CodeBuddy, Cline).
3. Tool Calling: Enables models to interact with external tools and APIs.
4. Web Search: Provides internet connectivity and information retrieval.
5. Multimodal Understanding: Includes capabilities for Image Understanding, Audio Understanding, and Video Understanding.
6. Speech Synthesis (TTS): Offered through the MiMo-V2.5-TTS Series (including mimo-v2.5-tts, mimo-v2.5-tts-voiceclone, mimo-v2.5-tts-voicedesign) and MiMo-V2-TTS, which allows for generating speech from text.
7. Agent Products: Supports multi-turn conversations with specific handling for reasoning_content to enhance agent performance.

Pricing Model and Billing Methodology:
The MiMo platform primarily operates on a Pay-As-You-Go (PAG) API Token Plan, distinct from any separate "Token Plan package quota." Billing is based on actual token usage, consuming from the user's account balance via a standard Open Platform API Key.

The billing unit varies by region:

• China: RMB (Chinese Yuan) per Million (M) tokens.
• Overseas: USD (United States Dollar) per Million (M) tokens.

Token consumption is differentiated by:

• Input Tokens:
  • Cache Hit: Billed at a reduced rate when the requested prompt prefix content matches data already present in the Prompt Cache. This optimization leverages previously processed or common input patterns to reduce costs and latency.
  • Cache Miss: Billed at the standard input rate for content not found in the cache.
• Output Tokens: Billed for the generated response content.

Internet search (Web Search) is billed independently per call, separate from token consumption. Cache Write operations are noted as "Limited-time Free."

Model Versioning and Deprecation:
The platform features two primary model series: MiMo-V2.5 and MiMo-V2. A price reduction for the MiMo-V2.5 series is scheduled to be effective on May 27, 2026, at 00:00 (GMT+8). The older MiMo-V2 models are slated for deprecation, with users advised to migrate to the newer V2.5 models.

Detailed Pricing Breakdown:

Domestic Pricing (RMB/M tokens):

• MiMo-V2.5 Series (Effective May 27, 2026):
  • mimo-v2.5-pro: Input (Cache Hit) ¥0.025; Input (Cache Miss) ¥3.00; Output ¥6.00.
  • mimo-v2.5: Input (Cache Hit) ¥0.02; Input (Cache Miss) ¥1.00; Output ¥2.00.
• MiMo-V2 Series (Pricing remains unchanged, models to be deprecated):
  • mimo-v2-pro:
    • Input $\leq 256\text{K}$: Input (Cache Hit) ¥1.40; Input (Cache Miss) ¥7.00; Output ¥21.00.
    • Input $256\text{K} - 1\text{M}$: Input (Cache Hit) ¥2.80; Input (Cache Miss) ¥14.00; Output ¥42.00.
  • mimo-v2-omni: Input $\leq 256\text{K}$: Input (Cache Hit) ¥0.56; Input (Cache Miss) ¥2.80; Output ¥14.00. (No pricing provided for Input $256\text{K} - 1\text{M}$).
  • off-v2-flash: Input $\leq 256\text{K}$: Input (Cache Hit) ¥0.07; Input (Cache Miss) ¥0.70; Output ¥2.10. (No pricing provided for Input $256\text{K} - 1\text{M}$).
• TTS Series: mimo-v2.5-tts, mimo-v2.5-tts-voiceclone, mimo-v2.5-tts-voicedesign, mimo-v2-tts are free for a limited time.

Overseas Pricing (USD/M tokens):

• MiMo-V2.5 Series:
  • mimo-v2.5-pro: Input (Cache Hit) $0.0036; Input (Cache Miss)$0.435; Output $0.87.
  • mimo-v2.5: Input (Cache Hit) $0.0028; Input (Cache Miss)$0.14; Output $0.28.
• MiMo-V2 Series:
  • mimo-v2-pro:
    • Input $\leq 256\text{K}$: Input (Cache Hit) $0.20; Input (Cache Miss)$1.00; Output $3.00.
    • Input $256\text{K} - 1\text{M}$: Input (Cache Hit) $0.40; Input (Cache Miss)$2.00; Output $6.00.
  • mimo-v2-omni: Input $\leq 256\text{K}$: Input (Cache Hit) $0.08; Input (Cache Miss)$0.40; Output $2.00. (No pricing provided for Input$256\text{K} - 1\text{M}$).
  • off-v2-flash: Input $\leq 256\text{K}$: Input (Cache Hit) $0.01; Input (Cache Miss)$0.10; Output $0.30. (No pricing provided for Input$256\text{K} - 1\text{M}$).
• TTS Series: mimo-v2.5-tts, mimo-v2.5-tts-voiceclone, mimo-v2.5-tts-voicedesign, mimo-v2-tts are free for a limited time.

Web Search Plugins Pricing:

• Domestic Internet Connectivity Service: ¥25 per 1000 calls (includes web search and parsing for domestic regions).
• Overseas Internet Connectivity Service: $5 per 1000 calls (includes web search and parsing for overseas regions).

Other Announcements:
The document also references news such as the conclusion of the "100 Trillion Token Creator Incentive Plan," the open-sourcing of Xiaomi MiMo-V2.5 series, and the launch of the Orbit 100 trillion token plan, indicating ongoing development and community engagement initiatives.