100 Tokens - AI Compute Power Package

📌 Package Overview

This package includes 100 Tokens of high-stability AI compute power, which can be used to call multiple mainstream AI models through One API redirection (all models are mapped to high-performance computing nodes with low latency).

✅ Supported Models

Core Supported (Low Multiplier Models)

• GPT-3.5-turbo series (all versions)
• ERNIE-Lite / ERNIE-Speed series
• SparkDesk full series (v1.1/v2.1/v3.1/v3.5/v4.0)
• glm-4-air / glm-4-flash / DeepSeek-chat
• Baichuan2-Turbo / Baichuan2-Turbo-192k

Compatible Supported (Medium Multiplier Models)

• GPT-4o (2024-11-20)
• Claude-3-5-Sonnet / Claude-3-Haiku
• ERNIE-3.5-8K series
• Moonshot-v1-8K / Moonshot-v1-32k

🔍 Token Deduction Rules

1 Token = 1000 calls of GPT-3.5-turbo (multiplier 0.25)

• GPT-3.5-turbo: 0.00025 Token per call
• GPT-4o (2024-11-20): 0.00125 Token per call
• Claude-3-5-Sonnet: 0.0015 Token per call
• ERNIE-3.5-8K: 0.000852 Token per call

🚫 Unsupported Models (Need Flagship Package)

• GPT-4 full series (including GPT-4 32K)
• Claude-3-Opus
• ERNIE-4.0-8K
• glm-4 / CogViewX

⚠️ Important Notes

1. All model calls are redirected to high-stability computing nodes via One API, with stable performance and low latency;
2. Valid for 30 days from the date of purchase, unused tokens will expire automatically;
3. After payment, you will receive the One API key instantly (auto-delivery), which can be used to query remaining tokens in real time;
4. No refund for manual violation of usage rules (e.g., malicious brushing of compute power).

📞 After-Sales Support

If you encounter compute power exceptions (e.g., call failure, incorrect token deduction), please contact our customer service within 24 hours for resolution.