feat(llm): add NVIDIA NIM, Kimi, Groq, and Mistral providers

Each is a one-row entry in the OpenAI-compatible provider registry (base_url,
key env, CLI option); the model is user-specified since they serve many models.

tradingagents/llm_clients/model_catalog.py

@@ -7,6 +7,13 @@ from typing import Dict, List, Tuple
 ModelOption = Tuple[str, str]
 ProviderModeOptions = Dict[str, Dict[str, List[ModelOption]]]
 
+# Providers that serve many / frequently-changing models: offer only "Custom
+# model ID" rather than a list that goes stale.
+_CUSTOM_ONLY: Dict[str, List[ModelOption]] = {
+    "quick": [("Custom model ID", "custom")],
+    "deep": [("Custom model ID", "custom")],
+}
+
 
 # Shared model list for GLM via Z.AI (international) and BigModel (China).
 # Source: docs.z.ai (GLM Coding Plan supported models + LLM guides).
@@ -177,10 +184,15 @@ MODEL_OPTIONS: ProviderModeOptions = {
     },
     # Generic OpenAI-compatible endpoint: the model is whatever the user's
     # server serves, so only "Custom model ID" is offered.
-    "openai_compatible": {
-        "quick": [("Custom model ID", "custom")],
-        "deep": [("Custom model ID", "custom")],
-    },
+    "openai_compatible": _CUSTOM_ONLY,
+    # Hosted OpenAI-compatible providers that serve many (and frequently
+    # changing) models — offer "Custom model ID" rather than a list that goes
+    # stale. The endpoint + key are wired by the provider; the user picks the
+    # model their account has access to.
+    "mistral": _CUSTOM_ONLY,
+    "kimi": _CUSTOM_ONLY,
+    "groq": _CUSTOM_ONLY,
+    "nvidia": _CUSTOM_ONLY,
 }