* feat: Add GPT-5 model family with official OpenAI specifications (#13378)
* Add GPT-5 model family support
Added four new GPT-5 models:
- gpt-5: Flagship model for logic and multi-step tasks
- gpt-5-mini: Cost-sensitive version for budget use cases
- gpt-5-nano: Speed-optimized version for low latency
- gpt-5-chat: Enterprise-focused version for advanced conversations
* Update GPT-5 models with official OpenAI specifications
- Add gpt-5-chat-latest with 400k context, 128k output tokens
- Add gpt-5-2025-08-07 with enhanced reasoning capabilities
- Add gpt-5-mini-2025-08-07 with cost-optimized pricing
- Add gpt-5-nano-2025-08-07 with ultra-fast performance
- Update existing gpt-5, gpt-5-mini, gpt-5-nano to match dated versions
- All models now support reasoning tokens and 400k context window
- Pricing updated per official OpenAI documentation
* fix conflicts
---------
Co-authored-by: Cole McIntosh <82463175+colesmcintosh@users.noreply.github.com>
Fixed MyPy type error in Ollama completion transformation where finish_reason
was set to None instead of expected string type. Changed finish_reason=None to
finish_reason="" to match GenericStreamingChunk TypedDict requirements.
Also updated corresponding test to expect empty string instead of None.
- Handle chunks containing 'thinking' field with empty 'response'
- Treat these as intermediate chunks that don't contain user content
- Add comprehensive tests for chunk parsing scenarios
- Resolves APIConnectionError for GPT-OSS model streaming
Fixes#13340
* (#13284) add avector_store_create to route_type which doesn't require model
* (#13284) exclude hidden params in metadata when create vector store
* (#13284) fix lint error
* (#13284) keep metadata None if metadata is None(not empty dict)
* (#13284) add test code
* (#13284) change test code name
* (#13284) add avector_store_search to route_type which doesn't require model
* feat: Add logo customization for LiteLLM admin UI
- Add simple logo customization page under Settings > UI Theme
- Support custom logo URL with instant preview
- Clean reset to default functionality
- Wrap app with ThemeProvider for proper context
- Backend properly handles null/empty logo URLs
* style: Update save button to use indigo color and 'Save Changes' text for consistency
* fix: Restore CloudZero background job initialization
Address PR review feedback by restoring the _initialize_spend_tracking_background_jobs method
that handles both CloudZero and Prometheus background job initialization. This ensures
critical spend tracking functionality is not lost while preserving the logo customization feature.
Changes:
- Restore _initialize_spend_tracking_background_jobs method
- Replace inline Prometheus initialization with proper method call
- Maintain CloudZero background job setup for cost tracking
* fix unsupported operand type(s) for +=: 'NoneType' and 'str' on clientside auth creds for responses
* fix the client side auth to use correct metadata
* add more tests
* fix tests
