Set the max tokens as the default is set to 4086

[iperezx]

No description provided.

[github-actions]

Code review by ChatGPT

[github-actions]

Adding max_tokens: 10000 seems reasonable, but consider:

• Cost: Increasing the token limit directly impacts the cost of API calls. Is 10000 tokens justified for the expected use cases? Could this lead to unexpected high costs?
• Context Length of Model: Verify that the underlying LLM supports a context window of at least 10000 tokens. If it doesn't, the request will likely be truncated (and potentially cause unexpected behavior) or fail outright. You should select a maximum token value that is compatible with the model.
• Default Value/Configuration: Is 10000 the right default? It's generally a good practice to allow configuring this parameter, perhaps via an environment variable or command-line argument, so users can tailor it to their needs and budget.
• Error Handling: It would be good to include some error handling in the process. For example check if the value of max_tokens is positive.