feat: improve R1 format setting labels to be more generic for reasoning models

Updated the "Enable R1 model parameters" setting label and description to be
more descriptive and inclusive of other reasoning/thinking models. The new
labeling explicitly mentions:
- DeepSeek R1
- Kimi-K2
- QWQ
- NVIDIA NIM and other OpenAI-compatible providers

This makes it clearer to users that the setting should be enabled when using
any reasoning/thinking model through OpenAI-compatible APIs, not just R1.

Addresses Issue #10969 (user suggestion from @bozoweed)

Author: roomote

webview-ui/src/i18n/locales/ca/settings.json

@@ -887,8 +887,8 @@
 		"cacheReadsPrice": "Cache reads price",
 		"cacheWritesPrice": "Cache writes price",
 		"enableStreaming": "Enable streaming",
-		"enableR1Format": "Enable R1 model parameters",
-		"enableR1FormatTips": "Must be enabled when using R1 models such as QWQ to prevent 400 errors",
+		"enableR1Format": "Enable reasoning model format",
+		"enableR1FormatTips": "Enable for reasoning/thinking models (DeepSeek R1, Kimi-K2, QWQ, etc.) to properly handle their response format. Required for models on NVIDIA NIM and other OpenAI-compatible providers.",
 		"useAzure": "Use Azure",
 		"azureApiVersion": "Set Azure API version",
 		"gemini": {