Update openai-models.md to fix 'Adding more OpenAI models' section (#572

)

fixes #333

docs/openai-models.md

@@ -106,9 +106,11 @@ Let's say OpenAI have just released the `gpt-3.5-turbo-0613` model and you want
 
 ```yaml
 - model_id: gpt-3.5-turbo-0613
+  model_name: gpt-3.5-turbo-0613
   aliases: ["0613"]
 ```
-The `model_id` is the identifier that will be recorded in the LLM logs. You can use this to specify the model, or you can optionally include a list of aliases for that model.
+
+the `model_id` is what LLM uses to identify the model in its logs. For the API, you’ll need the `model_name` -- this might be different from the `model_id`, especially if there's a chance of conflicts with other models. Use `model_id` to specify the model, and feel free to add aliases if you like.
 
 If the model is a completion model (such as `gpt-3.5-turbo-instruct`) add `completion: true` to the configuration.