Hi! Is there an API key that I can use while I develop my web app without waste credit in testing mode?
Thanks!
Welcome to the forum!
No.
Over a year ago OpenAI did give about $15 in free credits but that is no longer done.
Thanks for your reply, I use Gpt4-o for create a travel plan.
Can you tell me how call can I complete using Gpt4-o?
I mean, using 20$ dollars credit how can I simulate how many plans can I create? Thanks!
Hi, you could probably try to estimate how much you will spend using the pricing table for the api. (https://openai.com/api/pricing/)
For gpt4o these are the prices:
The only problem would be to guess how many tokens are given as input and returned as output… i guess you can try once and see how many tokens you used , but I’m sure someone could give you a better solution that I am not aware of.
Startup can apply for credits if ever needed!
Here is a rough guide,
First, just write out for yourself what you anticipate the conversation with the assistant should look like.
This will include,
- Your
system
message - Conversation pairs,
user
messageassistant
message
Example:
system
You are DadJokeGPT, an advanced large language model engineered to draft only the very best, new, novel, and never-before seen dad-jokes.
Users will provide to you a subject or topic and you will write an entirely original, extremely clever, and painfully punny dad-joke—a real “groaners!”
user
I would like a dad joke about pickleball.
assistant
You should never play pickleball with a cucumber—they’re sour losers!
Take this text[1] and either
- Pipe it through the
tiktoken
library - Put it into an online tokenizer like https://platform.openai.com/tokenizer
Here we can see that this toy exchange is 91-tokens, 77-tokens for input and 14-tokens for output
So, for this particular exchange the cost would be,
So, if this example accurately represents a typical use case, I would be able to get about 1,680 similar exchanges for $1.00.
You’ll just need to do the same steps for your particular use.
You’ll need to do the output separately so you can count the number of input and output tokens ↩︎
Thanks for all replies tu my question!!