[Bug]: OPENROUTER_MODEL defaults to paid model gpt-4o-mini — contradicts .env.example which documents a free model #289

Summary

The OPENROUTER_MODEL environment variable defaults to "openai/gpt-4o-mini" in backend/app/main.py (line 81), which is a paid model requiring OpenRouter credits. The .env.example file documents "nvidia/nemotron-3-super-120b-a12b:free" as the default, which is a free model. Users who only copy .env.example and set OPENROUTER_API_KEY without explicitly setting OPENROUTER_MODEL will get payment errors from OpenRouter.

Steps to reproduce

1. Copy .env.example to .env as instructed in README
2. Set OPENROUTER_API_KEY to a valid key
3. Do NOT set OPENROUTER_MODEL (expecting the documented free default)
4. Start the backend
5. Create an interview prep session
6. Observe: OpenRouter returns HTTP 402 (payment_required) because gpt-4o-mini requires credits

Expected behavior

The default model in code should match the documented default in .env.example. Since .env.example documents nvidia/nemotron-3-super-120b-a12b:free as the default, either:

• The code default should be changed to "nvidia/nemotron-3-super-120b-a12b:free", OR
• The .env.example should be updated to document "openai/gpt-4o-mini" and note that it requires paid credits

Actual behavior

The code defaults to "openai/gpt-4o-mini" (a paid model) while .env.example documents "nvidia/nemotron-3-super-120b-a12b:free" (a free model). New contributors following the setup guide who skip explicitly setting OPENROUTER_MODEL will encounter unexpected API errors.

Suggested fix

Change the fallback default in backend/app/main.py line 81 to match .env.example:

OPENROUTER_MODEL = os.getenv("OPENROUTER_MODEL", "nvidia/nemotron-3-super-120b-a12b:free")