feat: add llm_type and executor_type discriminators for checkpoint fidelity

Add type discriminator fields to BaseLLM subclasses and
BaseAgentExecutor subclasses so checkpoint deserialization restores
the correct provider class instead of always creating LLM/CrewAgentExecutor.

lib/crewai/src/crewai/agents/agent_builder/base_agent.py

@@ -64,11 +64,26 @@ def _serialize_crew_ref(value: Any) -> str | None:
     return str(value.id) if hasattr(value, "id") else str(value)
 
 
+_LLM_TYPE_REGISTRY: dict[str, str] = {
+    "base": "crewai.llms.base_llm.BaseLLM",
+    "litellm": "crewai.llm.LLM",
+    "openai": "crewai.llms.providers.openai.completion.OpenAICompletion",
+    "anthropic": "crewai.llms.providers.anthropic.completion.AnthropicCompletion",
+    "azure": "crewai.llms.providers.azure.completion.AzureCompletion",
+    "bedrock": "crewai.llms.providers.bedrock.completion.BedrockCompletion",
+    "gemini": "crewai.llms.providers.gemini.completion.GeminiCompletion",
+}
+
+
 def _validate_llm_ref(value: Any) -> Any:
     if isinstance(value, dict):
-        from crewai.llm import LLM
+        import importlib
 
-        return LLM(**value)
+        llm_type = value["llm_type"]
+        dotted = _LLM_TYPE_REGISTRY[llm_type]
+        mod_path, cls_name = dotted.rsplit(".", 1)
+        cls = getattr(importlib.import_module(mod_path), cls_name)
+        return cls(**value)
     return value
 
 
@@ -80,11 +95,22 @@ def _resolve_agent(value: Any, info: Any) -> Any:
     return Agent.model_validate(value, context=getattr(info, "context", None))
 
 
+_EXECUTOR_TYPE_REGISTRY: dict[str, str] = {
+    "base": "crewai.agents.agent_builder.base_agent_executor.BaseAgentExecutor",
+    "crew": "crewai.agents.crew_agent_executor.CrewAgentExecutor",
+    "experimental": "crewai.experimental.agent_executor.AgentExecutor",
+}
+
+
 def _validate_executor_ref(value: Any) -> Any:
     if isinstance(value, dict):
-        from crewai.agents.crew_agent_executor import CrewAgentExecutor
+        import importlib
 
-        return CrewAgentExecutor.model_validate(value)
+        executor_type = value["executor_type"]
+        dotted = _EXECUTOR_TYPE_REGISTRY[executor_type]
+        mod_path, cls_name = dotted.rsplit(".", 1)
+        cls = getattr(importlib.import_module(mod_path), cls_name)
+        return cls.model_validate(value)
     return value
 
 

lib/crewai/src/crewai/agents/agent_builder/base_agent_executor.py

@@ -21,6 +21,7 @@ if TYPE_CHECKING:
 class BaseAgentExecutor(BaseModel):
     model_config = {"arbitrary_types_allowed": True}
 
+    executor_type: str = "base"
     crew: Crew | None = Field(default=None, exclude=True)
     agent: BaseAgent | None = Field(default=None, exclude=True)
     task: Task | None = Field(default=None, exclude=True)

lib/crewai/src/crewai/agents/crew_agent_executor.py

@@ -96,6 +96,7 @@ class CrewAgentExecutor(BaseAgentExecutor):
     LLM interactions, tool execution, and feedback handling.
     """
 
+    executor_type: Literal["crew"] = "crew"
     llm: Annotated[
         BaseLLM | str | None,
         BeforeValidator(_validate_llm_ref),

lib/crewai/src/crewai/experimental/agent_executor.py

@@ -170,6 +170,7 @@ class AgentExecutor(Flow[AgentExecutorState], BaseAgentExecutor):  # type: ignor
 
     _skip_auto_memory: bool = True
 
+    executor_type: Literal["experimental"] = "experimental"
     suppress_flow_events: bool = True  # always suppress for executor
     llm: BaseLLM = Field(exclude=True)
     prompt: SystemPromptResult | StandardPromptResult = Field(exclude=True)

lib/crewai/src/crewai/llm.py

@@ -343,6 +343,7 @@ class AccumulatedToolArgs(BaseModel):
 
 
 class LLM(BaseLLM):
+    llm_type: Literal["litellm"] = "litellm"
     completion_cost: float | None = None
     timeout: float | int | None = None
     top_p: float | None = None

lib/crewai/src/crewai/llms/base_llm.py

@@ -117,6 +117,7 @@ class BaseLLM(BaseModel, ABC):
 
     model_config = ConfigDict(arbitrary_types_allowed=True, populate_by_name=True)
 
+    llm_type: str = "base"
     model: str
     temperature: float | None = None
     api_key: str | None = None

lib/crewai/src/crewai/llms/providers/anthropic/completion.py

@@ -148,6 +148,7 @@ class AnthropicCompletion(BaseLLM):
     offering native tool use, streaming support, and proper message formatting.
     """
 
+    llm_type: Literal["anthropic"] = "anthropic"
     model: str = "claude-3-5-sonnet-20241022"
     timeout: float | None = None
     max_retries: int = 2

lib/crewai/src/crewai/llms/providers/azure/completion.py

@@ -3,7 +3,7 @@ from __future__ import annotations
 import json
 import logging
 import os
-from typing import Any, TypedDict
+from typing import Any, Literal, TypedDict
 from urllib.parse import urlparse
 
 from pydantic import BaseModel, PrivateAttr, model_validator
@@ -74,6 +74,7 @@ class AzureCompletion(BaseLLM):
     offering native function calling, streaming support, and proper Azure authentication.
     """
 
+    llm_type: Literal["azure"] = "azure"
     endpoint: str | None = None
     api_version: str | None = None
     timeout: float | None = None

lib/crewai/src/crewai/llms/providers/bedrock/completion.py

@@ -5,7 +5,7 @@ from contextlib import AsyncExitStack
 import json
 import logging
 import os
-from typing import TYPE_CHECKING, Any, TypedDict, cast
+from typing import TYPE_CHECKING, Any, Literal, TypedDict, cast
 
 from pydantic import BaseModel, PrivateAttr, model_validator
 from typing_extensions import Required
@@ -228,6 +228,7 @@ class BedrockCompletion(BaseLLM):
     - Model-specific conversation format handling (e.g., Cohere requirements)
     """
 
+    llm_type: Literal["bedrock"] = "bedrock"
     model: str = "anthropic.claude-3-5-sonnet-20241022-v2:0"
     aws_access_key_id: str | None = None
     aws_secret_access_key: str | None = None

lib/crewai/src/crewai/llms/providers/gemini/completion.py

@@ -41,6 +41,7 @@ class GeminiCompletion(BaseLLM):
     offering native function calling, streaming support, and proper Gemini formatting.
     """
 
+    llm_type: Literal["gemini"] = "gemini"
     model: str = "gemini-2.0-flash-001"
     project: str | None = None
     location: str | None = None

lib/crewai/src/crewai/llms/providers/openai/completion.py

@@ -180,6 +180,8 @@ class OpenAICompletion(BaseLLM):
             chain-of-thought without storing data on OpenAI servers.
     """
 
+    llm_type: Literal["openai"] = "openai"
+
     BUILTIN_TOOL_TYPES: ClassVar[dict[str, str]] = {
         "web_search": "web_search_preview",
         "file_search": "file_search",