DeepSeek connector: add V4 models (deepseek-v4-pro / deepseek-v4-flash)

[leaft]

What

Add DeepSeek's current V4 models to the DeepSeek BYOK connector and make deepseek-v4-flash the recommended default, while keeping the existing deepseek-chat / deepseek-r1 options for backward compatibility.

-    defaultModel: 'deepseek-chat',
-    models: ['deepseek-chat', 'deepseek-r1'],
+    defaultModel: 'deepseek-v4-flash',
+    models: ['deepseek-v4-pro', 'deepseek-v4-flash', 'deepseek-chat', 'deepseek-r1'],

Why

The DeepSeek connector only offered the legacy deepseek-chat / deepseek-r1 models. Because the model picker for non-Ollama providers is a fixed <select> populated solely from provider.models, users had no way to select DeepSeek's current V4 models (deepseek-v4-pro, deepseek-v4-flash) per the official API docs.

Closes #103.

Verification

• DeepSeek API direct: POST https://api.deepseek.com/chat/completions with model: "deepseek-v4-pro" → HTTP 200, response echoes "model":"deepseek-v4-pro".
• Full lrc path (lrc ui → /api/ui/connectors/validate-key → LiveReview cloud → DeepSeek):
  • deepseek-v4-pro → {"valid":true}
  • deepseek-v4-flash (implicitly via default) — accepted
  • bogus model id → {"valid":false} (control: confirms the cloud actually probes the model, not just the key)
• go build ./... and node --check providers.js pass.

Scope

UI-only change to internal/staticserve/static/ui-connectors/providers.js; no backend changes.

[shrsv]

Hi @leaft - thanks for the contribution.

We will have a discussion about updating model names on a day to day basis from some source such as OpenRouter or such. This way - the outdated model problem can be eliminated in a general way.

If it's going to take a while to integrate that approach - we will merge your changes for the short term.

[LinceMathew]

Hi @leaft,

We discussed this issue in depth and came up with a permanent solution for keeping our models list up-to-date. Rather than updating the latest LLM models manually every time, we will fetch the models list from OpenRouter daily and keep our models up-to-date. I will roll out this feature today.

If this is in place, I don't think we need to merge these changes, since the latest DeepSeek models are also covered in the fix.

Thanks for pointing out this issue. Let us know if you face any other issues with the product.