ProxMenux/AppImage/config/verified_ai_models.json

{
  "_description": "Verified AI models for ProxMenux notifications. Only models listed here will be shown to users. Models are tested to work with the chat/completions API format.",
  "_updated": "2026-04-19",
  "_verifier": "Refreshed with tools/ai-models-verifier (private). Re-run before each ProxMenux release to keep the list current. The verifier and ProxMenux share the same reasoning/thinking-model handlers so their verdicts stay aligned with runtime behaviour.",

  "groq": {
    "models": [
      "llama-3.3-70b-versatile",
      "llama-3.1-70b-versatile",
      "llama-3.1-8b-instant",
      "llama3-70b-8192",
      "llama3-8b-8192",
      "mixtral-8x7b-32768",
      "gemma2-9b-it"
    ],
    "recommended": "llama-3.3-70b-versatile",
    "_note": "Not yet re-verified in 2026-04 refresh — kept from previous curation. Run the verifier with a Groq key to prune deprecated entries."
  },

  "gemini": {
    "models": [
      "gemini-2.5-flash-lite",
      "gemini-2.5-flash",
      "gemini-3-flash-preview"
    ],
    "recommended": "gemini-2.5-flash-lite",
    "_note": "flash-lite / flash pass the verifier consistently; pro variants reject thinkingBudget=0 and are overkill for notification translation anyway. 'latest' aliases (gemini-flash-latest, gemini-flash-lite-latest) are intentionally omitted because they resolved to different models across runs and produced timeouts in some regions.",
    "_deprecated": ["gemini-2.0-flash", "gemini-2.0-flash-lite", "gemini-1.5-flash", "gemini-1.0-pro", "gemini-pro"]
  },

  "openai": {
    "models": [
      "gpt-4.1-nano",
      "gpt-4.1-mini",
      "gpt-4o-mini",
      "gpt-4.1",
      "gpt-4o",
      "gpt-5-chat-latest",
      "gpt-5.4-nano",
      "gpt-5.4-mini"
    ],
    "recommended": "gpt-4.1-nano",
    "_note": "Reasoning models (o-series, gpt-5/5.1/5.2 non-chat variants) are supported by openai_provider.py via max_completion_tokens + reasoning_effort=minimal, but not listed here by default: their latency is higher than the chat models and they do not improve translation quality for notifications. Add specific reasoning IDs to this list only if a user explicitly wants them."
  },

  "anthropic": {
    "models": [
      "claude-3-5-haiku-latest",
      "claude-3-5-sonnet-latest",
      "claude-3-opus-latest"
    ],
    "recommended": "claude-3-5-haiku-latest",
    "_note": "Not re-verified in 2026-04 refresh — kept from previous curation. Add claude-4.x / claude-4.5 / claude-4.6 / claude-4.7 variants after running the verifier with an Anthropic key."
  },

  "openrouter": {
    "models": [
      "meta-llama/llama-3.3-70b-instruct",
      "meta-llama/llama-3.1-70b-instruct",
      "meta-llama/llama-3.1-8b-instruct",
      "anthropic/claude-3.5-haiku",
      "anthropic/claude-3.5-sonnet",
      "google/gemini-flash-1.5",
      "openai/gpt-4o-mini",
      "mistralai/mistral-7b-instruct",
      "mistralai/mixtral-8x7b-instruct"
    ],
    "recommended": "meta-llama/llama-3.3-70b-instruct",
    "_note": "Not re-verified in 2026-04 refresh. google/gemini-flash-2.5-flash-lite was malformed in the previous entry and has been replaced with google/gemini-flash-1.5."
  },

  "ollama": {
    "_note": "Ollama models are local, we don't filter them. User manages their own models.",
    "models": [],
    "recommended": ""
  }
}