FAQs
We trained a GPT-4 model to answer your questions based on our FAQs, Terms of Service, and Privacy Policy.
General FAQs
Frequently asked questions
Self-host FAQs
Frequently asked questions
Web Search FAQs
Token Usage FAQs
How tokens work?
- This is the total length of your current conversation. It is calculated by OpenAI's API when you send a message.
- When you reach a certain length, you cannot add any more messages to the chat. This is the context length limit of the OpenAI's API.
- GPT-3.5 allows maximum of 4,096 tokens per chat. GPT-4 allows maximum of 8,192 tokens per chat in the 8K model and 32,768 tokens in the 32K model. OpenAI may increase the context length limit in the future.
- Some enhanced features on TypingMind will increase your token usage. You can always turn these features off in the Settings pannel (click the gear icon in the sidebar). The enhanced features include: Chat Title Suggestion, Search Keywords Suggestions, AI Agents, Upload Document, etc. Amount of tokens used depends on the AI agent and the length of your document.
- If you reached the context length limit, you can delete some old messages to continue the chat, or start a new chat.
- Learn more about OpenAI's model, tokens, and context length here.
Cost Estimation FAQs
How is the cost calculated?
- Cost associated with a chat includes the token cost for the chat messages AND all of the enhanced features used for that chat.
- Some enhanced features on TypingMind will increase your token usage. You can always turn these features off in the Settings pannel (click the gear icon in the sidebar). The enhanced features include: Chat Title Suggestion, Search Keywords Suggestions, AI Agents, Upload Document, etc. Amount of tokens used depends on the AI agent and the length of your document.
- All costs are estimated, please refer to your OpenAI dashboard for the most accurate cost of your API key.
- The cost is calculated based on the public pricing of OpenAI's API. Each model has its own pricing, each type of tokens in each model also has its own pricing.
- Token usages are not recorded when streaming response is enabled. We will work on improving this soon.
- Learn more about OpenAI's model, tokens, and context length here.
