Add Avian as an LLM provider

[avianion]

Hey! I'd like to add Avian (avian.io) as a provider. It's an OpenAI-compatible inference API, so it slots in using the existing openai-completions implementation — no new provider file needed, just model definitions and the env key mapping.

I already have a working branch with the changes (types, models, env key, tests all passing). Happy to discuss the approach if there's anything you'd want done differently.