API returns empty model catalog; all modelIds 404 (auth OK)

[simoblue1806]

What happens

• GET /models/v2/list/models returns {"error":null,"output":[]} — an empty array — with no filter and for every task value tried.
• Every modelId 404s with "Model does not exist or has yet to be added to the Bytez catalog", including the documented quickstarts Qwen/Qwen3-4B (chat docs) and openai-community/gpt2 (get-started), and closed-source ids (openai/gpt-4o-mini, anthropic/claude-sonnet-4-5, google/gemini-2.0-flash).
• The OpenAI-compatible POST /v1/chat/completions route returns the same 404.

Auth is fine (control test)

• Valid key → HTTP 200; deliberately invalid key → HTTP 401 Unauthorized.
• GET /models/v2/list/tasks works and returns 36 tasks.

Environment

• Python SDK bytez 3.0.1 on Windows 11, and reproduced with raw requests calls.
• Date: 2026-08-05.

Two doc/API bugs hit along the way

1. The HTTP reference says Authorization: YOUR_KEY, but only authorization: Key YOUR_KEY works (the SDK sends the latter).
2. GET /models/v2/list/models?task=text-generation returns HTTP 500 Expected parameter(s): modelId instead of a filtered list.