fix: use supported image model (#567)

## Description
<!-- What does this PR do? -->


## PR Type

<!-- Delete the types that don't apply --!>

🆕 New Feature
🐛 Bug Fix
💅 Refactor
📚 Documentation
🚦 Infrastructure

## Relevant issues

<!-- e.g. "Fixes #123" -->

## Checklist
- [ ] I have added unit tests that prove my fix/feature works
- [ ] New and existing tests pass locally
- [ ] Documentation was updated where necessary
- [ ] I have read and followed the [contribution
guidelines](https://github.com/mozilla-ai/any-llm/blob/main/CONTRIBUTING.md)```

Author: njbrake

tests/conftest.py

@@ -79,7 +79,7 @@ def provider_model_map() -> dict[LLMProvider, str]:
 def provider_image_model_map(provider_model_map: dict[LLMProvider, str]) -> dict[LLMProvider, str]:
     return {
         **provider_model_map,
-        LLMProvider.WATSONX: "mistralai/pixtral-12b",
+        LLMProvider.WATSONX: "meta-llama/llama-guard-3-11b-vision",
         LLMProvider.SAMBANOVA: "Llama-4-Maverick-17B-128E-Instruct",
         LLMProvider.NEBIUS: "openai/gpt-oss-20b",
         LLMProvider.OPENROUTER: "mistralai/mistral-small-3.2-24b-instruct:free",