litellm - 💡(How to fix) Fix [Feature]: Add a typed exception for (This model isn't mapped yet) [provider=mistral] [1 pull requests]

Check for existing issues

• I have searched the existing issues and checked that my issue is not a duplicate.

The Feature

Problem Statement

Currently, when a model (with mistral provider) is requested but not mapped in LiteLLM, the system throws a generic Exception with a message similar to:

"This model isn't mapped yet"

Using a generic exception type makes it difficult for client code to:

Reliably detect this specific failure scenario Implement targeted error handling or retries Differentiate it from other runtime or configuration errors

To reproduce

prompt_tokens_cost_usd, completion_tokens_cost_usd = litellm.cost_per_token( "mistral-medium-3.5-128b", 100, 100 custom_llm_provider="mistral" )

By the time this issue is being taken care of it could be that "mistral-medium-3.5-128b" is mapped, so you can put any unsupported model instead.

Possible Enhancements

class ModelNotMappedError(Exception): """Raised when a requested model is not mapped in LiteLLM.""" pass

raise ModelNotMappedError("This model isn't mapped yet. model={}, custom_llm_provider={}. Add it here - https://github.com/BerriAI/litellm/blob/main/model_prices_and_context_window.json.".format( model, custom_llm_provider)

Motivation, pitch

• Improved error handling: Consumers can catch this specific exception instead of relying on string matching
• Cleaner integrations: Downstream services can implement fallback logic (e.g., auto-mapping, alternative models)
• Better observability: Easier to track and categorize configuration issues
• Consistency: Aligns with best practices of using typed exceptions instead of generic ones

What part of LiteLLM is this about?

SDK (litellm Python package)

LiteLLM is hiring a founding backend engineer, are you interested in joining us and shipping to all our users?

No

Twitter / LinkedIn details

https://www.linkedin.com/in/amir-ben-khadda-4938b3127/?locale=en