### Check for existing issues

- [x] I have searched the existing issues and checked that my issue is not a duplicate.

### The Feature

I would like LiteLLM proxy model config to support forced parameter overrides.

LiteLLM already supports default parameters and provider transformations, but enterprise operators sometimes need to enforce parameters even if the client sends different values.

Example use cases:

- Force `max_tokens` for a model to avoid upstream errors. - Force `reasoning_effort=none` for models/providers where reasoning causes instability or unexpected cost. - Strip or override unsupported provider parameters. - Enforce safety or compatibility defaults at the gateway layer.

Suggested config direction:

```yaml model_list: - model_name: my-model litellm_params: model: openai/my-upstream-model model_info: forced_params: max_tokens: 4096 reasoning_effort: none

### Motivation, pitch

In production AI gateways, some upstream providers or models require strict parameter control.

Default parameters are useful, but they do not prevent clients from overriding them. Operators sometimes need a stronger policy layer to keep traffic st…