Hi, is there any simple way of having a customGPT with instructions to refer to more than 20 files in its memory / knowledge base? What I want to use would need probably nearing 50 files. I just want to know if there is a simple no code way to do this.
#Custom GPT with more than 20 files
stray osprey
upper lynxBOT
Hi @stray osprey. I'm kapa.ai, your friendly support bot. I will try to answer your question by searching through the knowledge base. :wave:
Currently, OpenAI allows up to 20 files to be uploaded to a custom GPT for its knowledge base source (https://help.openai.com/en/articles/8843948-knowledge-in-gpts). Each file can be up to 512 MB in size and can contain up to 2,000,000 tokens source (https://community.openai.com/t/846493).
If you need to refer to more than 20 files, you might need to consolidate your information into fewer, larger files, as long as each file does not exceed the size and token limits.
Please note that the only way to manage the files attached to a GPT is using the GPT Builder UI, which is best for applications where context changes infrequently source (https://help.openai.com/en/articles/8843948-knowledge-in-gpts).
If you're encountering issues with the file size or token limit, you might need to reduce the size of your files or split your content into smaller parts source (https://community.openai.com/t/846493).
Unfortunately, there is no simple, no-code way to increase the limit of 20 files per custom GPT at this time.
Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:
stray osprey
@upper lynx Do you know any other methods for a customGPT to always refer to these files, and if the files are bigger than 512MB in size?
upper lynxBOT
Let me check the knowledge base for you. One moment! :books:
If your files are larger than 512MB, you will need to reduce their size to meet the current file size limit for GPT models source (https://help.openai.com/en/articles/8555545-file-uploads-faq). This could involve compressing the files, removing unnecessary data, or splitting the files into smaller parts.
As for referring to the files, you can provide explicit instructions for using knowledge files, including specifying file names, in your prompts to the GPT model source (https://help.openai.com/en/articles/9358033-key-guidelines-for-writing-instructions-for-custom-gpts#h_07bd5d91e2).
However, if you need to work with a large number of files or very large files, you might want to consider using the vector_store objects in the Assistants API. You can attach a maximum of 10,000 files to file_search using vector_store objects. Each file can be at most 512 MB in size and have a maximum of 5,000,000 tokens source (https://platform.openai.com/docs/assistants/deep-dive/thread-locks).
Please note that by default, the size of all the files uploaded in your project cannot exceed 100 GB, but you can reach out to OpenAI's support team to increase this limit source (https://platform.openai.com/docs/assistants/deep-dive/thread-locks).
Remember, the GPT model can only process text data. If your files contain images, only the text will be processed source (https://help.openai.com/en/articles/8843948-knowledge-in-gpts).
Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:
stray osprey
@upper lynx explain vector_store and is it possible to use it no code style
upper lynxBOT
Let me check the knowledge base for you. One moment! :books:
A vector_store is a collection of processed files that can be used by the file_search tool in OpenAI's API source (https://platform.openai.com/docs/api-reference). It contains various properties such as:
id: The identifier, which can be referenced in API endpoints.object: The object type, which is alwaysvector_store.created_at: The Unix timestamp (in seconds) for when the vector store was created.name: The name of the vector store.usage_bytes: The total number of bytes used by the files in the vector store.file_counts: The number of files that are currently being processed, have been successfully processed, have failed to process, were cancelled, and the total number of files.status: The status of the vector store, which can be eitherexpired,in_progress, orcompleted. A status ofcompletedindicates that the vector store is ready for use.expires_after: The expiration policy for a vector store.
However, as of now, there is no no-code way to use vector_store. It requires using OpenAI's API, which involves coding. You can use various methods such as getVectorStore, modifyVectorStore, createVectorStore, and listVectorStores to interact with vector_store source (https://platform.openai.com/docs/api-reference).
If you're not comfortable with coding, you might need to work with a developer or use a platform that provides a user-friendly interface for OpenAI's API.
Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:
stray osprey
@upper lynx Hi, an extension to the previous question. Let's say I made a customGPT. Then I allow my friends to use it. If they use the customGPT extensively will I have to pay extra?
upper lynxBOT
Let me check the knowledge base for you. One moment! :books:
No, you will not have to pay extra fees if members of your ChatGPT Enterprise or ChatGPT Edu workspace reach their usage limit. They will simply not be able to use the model until their usage limit resets, which typically happens in a week's time source (https://help.openai.com/en/articles/9855712-openai-o1-models-faq-chatgpt-enterprise-and-edu#h_bbadada4b0).
However, it's important to note that the usage of the model is tied to the account's usage limit. If the usage limit is exceeded, the model will not be accessible until the limit resets. It's also worth noting that the usage limit is not per model but per account. So, if you have multiple models, they all share the same usage limit.
If you want to avoid reaching the usage limit, you might want to monitor your usage and consider upgrading your plan if necessary. For the latest information on pricing, please visit OpenAI's pricing page (https://openai.com/pricing).
Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:
stray osprey
@upper lynx what I mean is, let's say i have my account. User A have their own account, User B have their own account, etc. Let's say there are 9999 users that like my customGPT and use it daily. Due to the high number of usage of that customGPT, will I be charged?
upper lynxBOT
Let me check the knowledge base for you. One moment! :books:
No, you will not be charged for other users using your custom GPT model. The usage of the model by other users is tied to their own individual usage limits and not yours. Each user's usage is tracked separately and they are responsible for their own usage costs source (https://help.openai.com/en/articles/9855712-openai-o1-models-faq-chatgpt-enterprise-and-edu#h_bbadada4b0).
However, it's important to note that if you share your custom GPT model with others, you should ensure that access is restricted to intended users only, especially when sensitive data is involved. OpenAI recommends utilizing the “Invite only” sharing setting for controlling access to your Custom GPTs source (https://community.openai.com/t/921609).
For the latest information on pricing, please visit OpenAI's pricing page (https://openai.com/pricing).
Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:
stray osprey
@upper lynx so that means even if the whole world uses my customGPT, i wouldn't be charged a single cent. Correct?
upper lynxBOT
Let me check the knowledge base for you. One moment! :books:
Yes, that's correct. As long as the users are using their own accounts to access your custom GPT model, their usage is tied to their own individual usage limits and not yours. You will not be charged for their usage of your custom GPT model source (https://help.openai.com/en/articles/9855712-openai-o1-models-faq-chatgpt-enterprise-and-edu#h_bbadada4b0).
Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:
stray osprey
@upper lynx is the upload file section of the create customGPT disabled?
upper lynxBOT
Let me check the knowledge base for you. One moment! :books:
Based on the provided knowledge sources, there is no specific information about the upload file section of the custom GPT creation being disabled. However, there have been instances where users have reported issues with the user interface or specific features source (https://community.openai.com/t/781396), source (https://community.openai.com/t/759073).
If you're experiencing issues, it's recommended to check the OpenAI community forum or OpenAI's help center for any updates or ongoing issues. If the problem persists, you may want to reach out to OpenAI's support for assistance.
Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:
stray osprey
@upper lynx how many characters can a customGPT instruction fit
upper lynxBOT
Let me check the knowledge base for you. One moment! :books:
Each response in a custom instruction for ChatGPT has a limit of 1500 characters source (https://help.openai.com/en/articles/8096356-custom-instructions-for-chatgpt#h_5d941e87fa). This means that the instructions you provide to guide the behavior of your custom GPT model should not exceed this limit.
Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:
stray osprey
@upper lynx no, what I mean is when i create GPT, i can create / configure. In the create window, how many characters of instructions can the custom GPT hold
upper lynxBOT
Let me check the knowledge base for you. One moment! :books:
The system instructions that the assistant uses in the Assistants API has a maximum length of 256,000 characters source (https://help.openai.com/en/articles/8550641-assistants-api-v2-faq#h_c8ff59051b). This means that when you're creating a custom GPT model, the instructions you provide should not exceed this limit.
Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:
stray osprey
@upper lynx that is for assistant API. I am not building an assistant API. I am building a custom chatgpt based on the tool chatgpt gives us users.
upper lynxBOT
Let me check the knowledge base for you. One moment! :books: