@kauabh I agree that if a model API has a method to count tokens, it would be nice to expose that on theModel class.

But I don't think we should automatically use it whenUsageLimits(request_tokens_limit=...) is used, as it adds an extra request and the overhead and latency that comes with that, unlike OpenAI'stiktoken which was mentioned in#1794 and can be run locally. So if we'd like to give users the option to better enforcerequest_tokens_limit by doing a separate count-tokens request ahead of the actual LLM request, that should be opt-in with some flag onUsageLimits and appropriate warnings in the docs about the extra overhead.

That check would need to be implemented here, just before we callmodel.request, once we have the messages, model settings, and model request params ready:

This would require a method that exists on every model, so it'd be implemented as an abstract method on the baseModel class with a default implementation ofraise NotImplementedError(...), and only models that have a count-tokens method would override it with a concrete implementation.

As for that concrete implementation, I recommend adding it toGoogleModel instead ofGeminiModel, as you can directly use thegoogle-genai library there, and reducing the duplication with the request-preparation logic in_generate_content as much as possible.

@DouweM make sense, let me rework on this. Thanks for detailed input, appreciate your time