Update default reasoning model to GPT-5

AGENTS.md

@@ -26,7 +26,7 @@ Please adhere to the following code style guidelines:
 
 ## General Guidance
 
-*   The default model for this project is `gpt-4o`.
+*   The default model for this project is `gpt-5`.
 *   Write clear and descriptive commit messages.
 *   Break down large changes into smaller, logical commits.
 *   Ensure that all new code is covered by tests.

docs/api/core.md

@@ -13,14 +13,14 @@ class Agent:
     def __init__(
         self, 
         name: str, 
-        model: str = "gpt-4o",
+        model: str = "gpt-5",
         memory: Optional[Memory] = None
     ):
         """Initialize an agent with specified configuration.
 
         Args:
             name: Unique identifier for the agent
-            model: LLM model to use (default: "gpt-4o")
+            model: LLM model to use (default: "gpt-5")
             memory: Optional memory instance for conversation history
         """
 ```
@@ -259,7 +259,7 @@ class AgentConfig(BaseModel):
     model_config = ConfigDict(arbitrary_types_allowed=True)
 
     name: str
-    model: str = "gpt-4o"
+    model: str = "gpt-5"
     temperature: float = Field(default=0.7, ge=0.0, le=2.0)
     max_tokens: int = Field(default=1000, gt=0)
     memory_enabled: bool = True

docs/api/services.md

@@ -53,7 +53,7 @@ Calculates optimal staking APR using market data and sentiment analysis.
 
 ```python
 class YieldManagerService(Service):
-    def __init__(self, model: str = "gpt-4o"):
+    def __init__(self, model: str = "gpt-5"):
         """Initialize yield manager service.
 
         Args:

docs/architecture/agents.md

@@ -37,15 +37,15 @@ class MainAgent:
 The `Agent` class provides core functionality inherited by all specialized agents:
 
 **Core Features:**
-- **LLM Interaction** - Standardized interface to language models (default: GPT-4o)
+- **LLM Interaction** - Standardized interface to language models (default: GPT-5)
 - **Conversation History** - Maintains context across interactions using message history
 - **Memory Integration** - Semantic search and retrieval of past conversations
 - **Prompt Management** - Template-based prompt system with dynamic loading
 
 **Implementation Details:**
 ```python
 class Agent:
-    def __init__(self, model: str = "gpt-4o"):
+    def __init__(self, model: str = "gpt-5"):
         self.model = model
         self.history = []
         self.memory = Memory()
@@ -153,7 +153,7 @@ Agents can be configured with different LLM models:
 
 ```python
 # Default configuration
-agent = Agent(model="gpt-4o")
+agent = Agent(model="gpt-5")
 
 # Custom model for specific tasks
 code_review_agent = Agent(model="gpt-4o-code")

docs/architecture/hypervisor.md

@@ -248,7 +248,7 @@ approval_settings:
 ```yaml
 hypervisor:
   supervisor: "hybrid"
-  llm_model: "gpt-4o"
+  llm_model: "gpt-5"
   confidence_threshold: 0.8
 
 risk_categories:

docs/cli/proposals.md

@@ -16,7 +16,7 @@ uv run talos proposals eval --file <filepath>
 - `--file, -f`: Path to the proposal file (required)
 
 **Options:**
-- `--model-name`: LLM model to use (default: "gpt-4o")
+- `--model-name`: LLM model to use (default: "gpt-5")
 - `--temperature`: Temperature for LLM generation (default: 0.0)
 
 ## Usage Examples
@@ -28,7 +28,7 @@ uv run talos proposals eval --file <filepath>
 uv run talos proposals eval --file governance_proposal.txt
 
 # Use a different model
-uv run talos proposals eval --file proposal.md --model-name gpt-4o --temperature 0.1
+uv run talos proposals eval --file proposal.md --model-name gpt-5 --temperature 0.1
 ```
 
 ### Proposal File Format

docs/development/code-style.md

@@ -84,7 +84,7 @@ Use Google-style docstrings for all modules, classes, and functions:
 ```python
 def analyze_sentiment(
     text: str, 
-    model: str = "gpt-4o",
+    model: str = "gpt-5",
     confidence_threshold: float = 0.7
 ) -> SentimentResult:
     """Analyze sentiment of the given text using an LLM.
@@ -94,7 +94,7 @@ def analyze_sentiment(
 
     Args:
         text: The text to analyze for sentiment. Must not be empty.
-        model: The LLM model to use for analysis. Defaults to "gpt-4o".
+        model: The LLM model to use for analysis. Defaults to "gpt-5".
         confidence_threshold: Minimum confidence score to return results.
             Must be between 0.0 and 1.0.
 
@@ -137,7 +137,7 @@ class AgentConfig(BaseModel):
         extra='forbid'
     )
 
-    model_name: str = Field(default="gpt-4o", description="LLM model to use")
+    model_name: str = Field(default="gpt-5", description="LLM model to use")
     temperature: float = Field(default=0.7, ge=0.0, le=2.0)
     max_tokens: int = Field(default=1000, gt=0)
 ```
@@ -151,7 +151,7 @@ class Agent(BaseModel):
     model_config = ConfigDict(arbitrary_types_allowed=True)
 
     name: str
-    model: str = "gpt-4o"
+    model: str = "gpt-5"
     memory: Memory | None = None
 
     def model_post_init(self, __context: Any) -> None:
@@ -384,7 +384,7 @@ def mock_openai_client():
 def test_agent(mock_openai_client):
     """Create test agent with mocked dependencies."""
     with patch('talos.core.agent.OpenAI', return_value=mock_openai_client):
-        return Agent(name="test_agent", model="gpt-4o")
+        return Agent(name="test_agent", model="gpt-5")
 
 def test_agent_generates_response(test_agent):
     """Test that agent generates appropriate response."""

docs/development/contributing.md

@@ -197,7 +197,7 @@ from talos.utils.helpers import format_response
 ```python
 def test_agent_processes_query_successfully():
     # Arrange
-    agent = Agent(model="gpt-4o")
+    agent = Agent(model="gpt-5")
     query = "What is the current market sentiment?"
 
     # Act
@@ -239,7 +239,7 @@ uv run pytest -k "test_sentiment"
 Use Google-style docstrings:
 
 ```python
-def analyze_sentiment(text: str, model: str = "gpt-4o") -> float:
+def analyze_sentiment(text: str, model: str = "gpt-5") -> float:
     """Analyze sentiment of the given text.
 
     Args:

docs/development/testing.md

@@ -108,7 +108,7 @@ Follow the Arrange-Act-Assert pattern:
 ```python
 def test_agent_processes_query_successfully():
     # Arrange - Set up test data and dependencies
-    agent = Agent(name="test_agent", model="gpt-4o")
+    agent = Agent(name="test_agent", model="gpt-5")
     query = "What is the current market sentiment?"
     expected_response_type = QueryResponse
 
@@ -477,7 +477,7 @@ from talos.core.memory import Memory
 def test_config():
     """Test configuration for all tests."""
     return {
-        "model": "gpt-4o",
+        "model": "gpt-5",
         "test_mode": True,
         "api_timeout": 30
     }
@@ -498,7 +498,7 @@ def mock_api_keys(monkeypatch):
 @pytest.fixture
 def test_agent(mock_api_keys):
     """Create test agent with mocked dependencies."""
-    return Agent(name="test_agent", model="gpt-4o")
+    return Agent(name="test_agent", model="gpt-5")
 
 @pytest.fixture
 def test_memory(temp_directory):

examples/proposal_example.py

@@ -11,7 +11,7 @@ def run_proposal_example():
     Runs an example of the agent evaluating a proposal.
     """
     # Initialize the proposals service
-    service = ProposalsService(llm=ChatOpenAI(model="gpt-4o", openai_api_key=os.environ.get("OPENAI_API_KEY", "")))
+    service = ProposalsService(llm=ChatOpenAI(model="gpt-5", openai_api_key=os.environ.get("OPENAI_API_KEY", "")))
 
     # Define the proposal
     proposal = Proposal(

integration_tests/test_memory_integration.py

@@ -26,7 +26,7 @@ def teardown_method(self):
 
     def test_main_agent_memory_tool_registration(self):
         """Test that MainAgent properly registers memory tools."""
-        model = ChatOpenAI(model="gpt-4o", api_key="dummy-key")
+        model = ChatOpenAI(model="gpt-5", api_key="dummy-key")
 
         agent = MainAgent(
             model=model,
@@ -42,7 +42,7 @@ def test_main_agent_memory_tool_registration(self):
 
     def test_memory_storage_and_retrieval(self):
         """Test memory storage and retrieval functionality."""
-        model = ChatOpenAI(model="gpt-4o", api_key="dummy-key")
+        model = ChatOpenAI(model="gpt-5", api_key="dummy-key")
 
         agent = MainAgent(
             model=model,

integration_tests/test_memory_prompt_fix.py

@@ -21,7 +21,7 @@ def test_memory_tool_invocation():
     try:
         from langchain_openai import ChatOpenAI
 
-        model = ChatOpenAI(model="gpt-4o", api_key="dummy-key")
+        model = ChatOpenAI(model="gpt-5", api_key="dummy-key")
 
         agent = MainAgent(
             model=model,

integration_tests/test_memory_tool_availability.py

@@ -20,7 +20,7 @@ def test_memory_tool_availability():
     memory_file = Path(temp_dir) / "test_memory.json"
 
     try:
-        model = ChatOpenAI(model="gpt-4o", api_key="dummy-key")
+        model = ChatOpenAI(model="gpt-5", api_key="dummy-key")
 
         agent = MainAgent(
             model=model,

src/talos/cli/README.md

@@ -170,7 +170,7 @@ Run the Talos agent in daemon mode for continuous operation.
 
 ```bash
 uv run talos daemon
-uv run talos daemon --model-name gpt-4o --temperature 0.1
+uv run talos daemon --model-name gpt-5 --temperature 0.1
 ```
 
 #### `cleanup-users [--older-than] [--dry-run]`

src/talos/cli/daemon.py

@@ -16,7 +16,7 @@
 
 
 class TalosDaemon:
-    def __init__(self, prompts_dir: str = "src/talos/prompts", model_name: str = "gpt-4", temperature: float = 0.0):
+    def __init__(self, prompts_dir: str = "src/talos/prompts", model_name: str = "gpt-5", temperature: float = 0.0):
         self.prompts_dir = prompts_dir
         self.model_name = model_name
         self.temperature = temperature

src/talos/cli/main.py

@@ -54,7 +54,7 @@ def callback(
 def main_command(
     query: Optional[str] = typer.Argument(None, help="The query to send to the agent."),
     prompts_dir: str = "src/talos/prompts",
-    model_name: str = "gpt-4o",
+    model_name: str = "gpt-5",
     temperature: float = 0.0,
     verbose: int = typer.Option(
         0, "--verbose", "-v", count=True, help="Enable verbose output. Use -v for basic, -vv for detailed."
@@ -156,7 +156,7 @@ def decrypt(encrypted_data: str, key_dir: str = ".keys"):
 @app.command()
 def daemon(
     prompts_dir: str = "src/talos/prompts",
-    model_name: str = "gpt-4o",
+    model_name: str = "gpt-5",
     temperature: float = 0.0,
 ) -> None:
     """

src/talos/cli/proposals.py

@@ -10,7 +10,7 @@
 @proposals_app.command("eval")
 def eval_proposal(
     filepath: str = typer.Option(..., "--file", "-f", help="Path to the proposal file."),
-    model_name: str = "gpt-4o",
+    model_name: str = "gpt-5",
     temperature: float = 0.0,
 ):
     """

src/talos/core/memory.py

@@ -103,7 +103,7 @@ def _setup_langmem_sqlite(self):
             )
 
             self._langmem_manager = create_memory_store_manager(
-                "gpt-4o",
+                "gpt-5",
                 namespace=("memories", self.user_id or "default"),
                 store=self._store
             )
@@ -128,7 +128,7 @@ def _setup_langmem_file(self):
             return
 
         try:
-            self._langmem_manager = create_memory_manager("gpt-4o")
+            self._langmem_manager = create_memory_manager("gpt-5")
             if self.file_path:
                 self.file_path.parent.mkdir(parents=True, exist_ok=True)
                 if not self.file_path.exists():

src/talos/skills/twitter_persona.py

@@ -23,7 +23,7 @@ class TwitterPersonaSkill(Skill):
     model_config = ConfigDict(arbitrary_types_allowed=True)
     twitter_client: TwitterClient = Field(default_factory=TweepyClient)
     prompt_manager: PromptManager = Field(default_factory=lambda: FilePromptManager("src/talos/prompts"))
-    llm: Any = Field(default_factory=lambda: ChatOpenAI(model="gpt-4o"))
+    llm: Any = Field(default_factory=lambda: ChatOpenAI(model="gpt-5"))
 
     @property
     def name(self) -> str:

src/talos/skills/twitter_voice.py

@@ -20,7 +20,7 @@ class TwitterVoiceSkill(Skill):
 
     model_config = ConfigDict(arbitrary_types_allowed=True)
     prompt_manager: PromptManager = Field(default_factory=lambda: FilePromptManager("src/talos/prompts"))
-    llm: Any = Field(default_factory=lambda: ChatOpenAI(model="gpt-4o"))
+    llm: Any = Field(default_factory=lambda: ChatOpenAI(model="gpt-5"))
     twitter_persona_skill: TwitterPersonaSkill | None = Field(default=None)
 
     @property

tests/test_scheduled_jobs.py

@@ -174,7 +174,7 @@ def main_agent(self):
             'OPENAI_API_KEY': 'test_openai_key'
         }):
             agent = MainAgent(
-                model=ChatOpenAI(model="gpt-4o", api_key="test_key"),
+                model=ChatOpenAI(model="gpt-5", api_key="test_key"),
                 prompts_dir="src/talos/prompts"
             )
             if agent.job_scheduler:
@@ -221,7 +221,7 @@ def test_predefined_jobs_registration(self):
             'OPENAI_API_KEY': 'test_openai_key'
         }):
             agent = MainAgent(
-                model=ChatOpenAI(model="gpt-4o", api_key="test_key"),
+                model=ChatOpenAI(model="gpt-5", api_key="test_key"),
                 prompts_dir="src/talos/prompts",
                 scheduled_jobs=[job]
             )