[FEATURE]: Add deepseek-v4-flash-vision-exp to DeepSeek provider

[Haicaji]

Feature hasn't been suggested before.

• I have verified this feature hasn't been suggested before.

Category

Provider / Model

Description

Add deepseek-v4-flash-vision-exp to DeepSeek provider to support multimodal capabilities.

[badchars]

Hi @Haicaji, thanks for taking the time to open this issue! 🙏

Great news — deepseek-v4-flash-vision-exp is actually already supported. All DeepSeek models (including Chat, V4 Pro, V4 Flash, V4 Flash Vision Exp, and Reasoner) are available once you connect your DeepSeek API key.

To set it up:

1. Via the TUI: Press Ctrl+A → "Connect provider" → select DeepSeek → enter your API key
2. Via CLI: cyberstrike auth login deepseek
3. Via environment variable: export DEEPSEEK_API_KEY="sk-..."

Once connected, type /models and search for "deep" — you'll see all available DeepSeek models including the vision model.

CyberStrike uses a dynamic model catalog, so provider models only appear in the model picker after authentication. This is by design — we only show models you can actually use.

Closing this as already supported, but please don't hesitate to open new issues if you run into any bugs, have feedback on the overall experience, or want to suggest other improvements. Community feedback is incredibly valuable to us and helps shape the direction of the project. We'd love to hear how things work for you!