AI Credits Consumption Rules
1. Time-limited credits will be deducted first, followed by permanent credits.
2. Credits included in your package will automatically refresh on your monthly purchase date and do not accumulate.
3. Using AI features consumes credits. The amount consumed depends on the specific feature, model, and word count.
4. Different AI models use varying **Token** calculation methods for text. The exact rules are determined by each model provider. The input and output **Tokens** consumed are directly returned by the model and are beyond our control. If you have any questions about credit deductions, please contact our online customer service.
5. The AI models used for various features may change. Each model has a different text **Token** calculation method, so the following consumption rules are for reference only.
6. For chain-of-thought models, the thinking process is also part of the output, which will count tokens and consume credits.
AI Feature Credit Consumption Rules
Applicable to English
| Feature | Basic Model | Advanced Model | Description | Remarks |
|---|---|---|---|---|
| AI Chat/Long Text Summary/File Summary/Link Summary | Input 1000 tokens = 1 credit (Less than 1000 tokens consumed proportionally) Output 1000 tokens = 4 credits | Input 1000 tokens = 6 credits (Less than 1000 tokens consumed proportionally) Output 1000 tokens = 24 credits | Input + Output = Total credits deducted (rounded) Example: | The following Token calculation rules are for reference only. The specific rules are determined by each model provider.
|
| Web Search (Requires reading multiple web pages, higher credit consumption) | Input 1000 tokens = 1 credit (Less than 1000 tokens consumed proportionally) Output 1000 tokens = 6 credits | Input 1000 tokens = 6 credits (Less than 1000 tokens consumed proportionally) Output 1000 tokens = 26 credits | ||
| Image Summary | Analyze 1 image = 20 credits Output 1000 tokens = 8 credits (Less than 1000 tokens consumed proportionally) | Example: Input one image=20 credits Output 100 tokens ≈ 0.8 credits Total credits deducted (rounded) = 21 credits | ||
| Audio Summary | 4 credits deducted per minute, less than one minute is counted as one minute | Example: 5 minutes of audio = 20 credits | ||
| Youtube Video Summary | If the video has subtitles, 1 credit is consumed per minute. If the video does not have subtitles, 4 credits are consumed per minute. If it is less than one minute, it will be counted as one minute. | Example: 10 minutes of video without subtitles = 40 credits | ||
| AI Art | Quality + Ultra HD = 200 credits Quality + HD = 80 credits Speed + Ultra HD = 180 credits Speed + HD = 60 credits | |||
Additional Notes:
Currently, most GitMind AI features utilize ChatGPT models. Since OpenAI charges based on Tokens, using non-English languages may result in higher usage costs due to a greater number of Tokens. Below is an explanation from OpenAI’s official documentation:
https://help.openai.com/en/articles/4936856-what-are-tokens-and-how-to-count-them
GitMind AI Models and Token Calculation
GitMind currently uses the following models. For detailed Token calculation explanations, please refer to the official documentation of each large language model:
OpenAI
- Tokenizer Tool (directly test text Token count): https://platform.openai.com/tokenizer
- Official Documentation (Token Calculation Explanation): https://platform.openai.com/docs/guides/text-generation/managing-tokens
Leave a Comment