Support image tokenizer for GPT-4.1 family of models

As per [the doc](https://platform.openai.com/docs/guides/images-vision?api-mode=responses#gpt-4-1-mini-gpt-4-1-nano-o4-mini)

Note that Azure [documentation](https://learn.microsoft.com/en-us/azure/ai-foundry/openai/overview#image-input-tokens) is currently missing information about image tokens for GPT-4.1 and o-series models.