feat(llm): Add Kimi K2 Thinking model to OpenRouter client

Author: lzwjava

scripts/llm/openrouter_client.py

@@ -21,7 +21,8 @@
     "grok-code": "x-ai/grok-code-fast-1",
     "grok-4-fast": "x-ai/grok-4-fast",
     "glm-4.6": "z-ai/glm-4.6",
-    "minimax-m2": "minimax/minimax-m2:free"
+    "minimax-m2": "minimax/minimax-m2:free",
+    "kimi-k2-thinking": "moonshotai/kimi-k2-thinking"
 }
 
 DEFAULT_TOKENS = {
@@ -37,7 +38,8 @@
     "grok-code": 62144,
     "grok-4-fast": 61072,
     "glm-4.6": 32768,
-    "mimimax-m2": 32768
+    "mimimax-m2": 32768,
+    "kimi-k2-thinking": 32768    
 }