Reduce model temperatures to 0.

CHANGELOG.md

@@ -13,6 +13,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Added ask_human tool to allow human operator to answer questions asked by the agent.
 - Handle keyboard interrupt (ctrl-c.)
 - Disable PAGERs for shell commands so agent can work autonomously.
+- Reduce model temperatures to 0.
 
 ## [0.6.4] - 2024-12-19
 

ra_aid/llm.py

@@ -22,24 +22,28 @@ def initialize_llm(provider: str, model_name: str) -> BaseChatModel:
     if provider == "openai":
         return ChatOpenAI(
             api_key=os.getenv("OPENAI_API_KEY"),
-            model=model_name
+            model=model_name,
+            temperature=0
         )
     elif provider == "anthropic":
         return ChatAnthropic(
             api_key=os.getenv("ANTHROPIC_API_KEY"),
-            model_name=model_name
+            model_name=model_name,
+            temperature=0
         )
     elif provider == "openrouter":
         return ChatOpenAI(
             api_key=os.getenv("OPENROUTER_API_KEY"),
             base_url="https://openrouter.ai/api/v1",
-            model=model_name
+            model=model_name,
+            temperature=0
         )
     elif provider == "openai-compatible":
         return ChatOpenAI(
             api_key=os.getenv("OPENAI_API_KEY"),
             base_url=os.getenv("OPENAI_API_BASE"),
-            model=model_name
+            model=model_name,
+            temperature=0
         )
     else:
         raise ValueError(f"Unsupported provider: {provider}")
@@ -64,25 +68,29 @@ def initialize_expert_llm(provider: str = "openai", model_name: str = "o1-previe
     if provider == "openai":
         return ChatOpenAI(
             api_key=os.getenv("EXPERT_OPENAI_API_KEY"),
-            model=model_name
+            model=model_name,
+            temperature=0
         )
     elif provider == "anthropic":
         return ChatAnthropic(
             api_key=os.getenv("EXPERT_ANTHROPIC_API_KEY"),
             model_name=model_name,
-            max_tokens=200000
+            max_tokens=200000,
+            temperature=0
         )
     elif provider == "openrouter":
         return ChatOpenAI(
             api_key=os.getenv("EXPERT_OPENROUTER_API_KEY"),
             base_url="https://openrouter.ai/api/v1",
-            model=model_name
+            model=model_name,
+            temperature=0
         )
     elif provider == "openai-compatible":
         return ChatOpenAI(
             api_key=os.getenv("EXPERT_OPENAI_API_KEY"),
             base_url=os.getenv("EXPERT_OPENAI_API_BASE"),
-            model=model_name
+            model=model_name,
+            temperature=0
         )
     else:
         raise ValueError(f"Unsupported provider: {provider}")