diff --git a/sdks/python/agenta/sdk/utils/types.py b/sdks/python/agenta/sdk/utils/types.py
index 8e629b92fb..01bc5003ad 100644
--- a/sdks/python/agenta/sdk/utils/types.py
+++ b/sdks/python/agenta/sdk/utils/types.py
@@ -2,7 +2,17 @@
 from copy import deepcopy
 from dataclasses import dataclass
 from enum import Enum
-from typing import Annotated, ClassVar, List, Union, Optional, Dict, Literal, Any
+from typing import (
+    Annotated,
+    ClassVar,
+    List,
+    Union,
+    Optional,
+    Dict,
+    Literal,
+    Any,
+    TypeAlias,
+)
 
 from pydantic import ConfigDict, BaseModel, HttpUrl, RootModel
 from pydantic import Field, model_validator, AliasChoices
@@ -406,6 +416,33 @@ class ResponseFormatJSONSchema(BaseModel):
 ]
 
 
+class ChatCompletionNamedToolChoiceFunction(BaseModel):
+    name: str
+
+
+class ChatCompletionNamedToolChoice(BaseModel):
+    type: Literal["function"]
+    function: ChatCompletionNamedToolChoiceFunction
+
+
+class ChatCompletionAllowedTools(BaseModel):
+    mode: Literal["auto", "required"]
+    tools: List[Dict[str, Any]]
+
+
+class ChatCompletionAllowedToolChoice(BaseModel):
+    type: Literal["allowed_tools"]
+    allowed_tools: ChatCompletionAllowedTools
+
+
+ToolChoice: TypeAlias = Union[
+    Literal["none", "auto", "required"],
+    ChatCompletionNamedToolChoice,
+    ChatCompletionAllowedToolChoice,
+    Dict[str, Any],
+]
+
+
 class ModelConfig(BaseModel):
     """Configuration for model parameters"""
 
@@ -468,7 +505,7 @@ class ModelConfig(BaseModel):
         default=None,
         description="A list of tools the model may call. Currently, only functions are supported as a tool",
     )
-    tool_choice: Optional[Union[Literal["none", "auto"], Dict]] = Field(
+    tool_choice: Optional[ToolChoice] = Field(
         default=None, description="Controls which (if any) tool is called by the model"
     )
 
@@ -596,7 +633,7 @@ class AgLLM(AgSchemaMixin):
         json_schema_extra={"x-ag-type": "choice"},
     )
     chat_template_kwargs: Optional[Dict[str, Any]] = Field(default=None)
-    tool_choice: Optional[Union[Literal["none", "auto"], Dict]] = Field(
+    tool_choice: Optional[ToolChoice] = Field(
         default=None,
     )
     template_format: Literal["mustache", "curly", "fstring", "jinja2"] = Field(
@@ -1015,7 +1052,13 @@ def to_openai_kwargs(self, llm_config: Optional[ModelConfig] = None) -> dict:
             kwargs["tools"] = llm_config.tools
             # Only set tool_choice if tools are present
             if llm_config.tool_choice is not None:
-                kwargs["tool_choice"] = llm_config.tool_choice
+                if isinstance(llm_config.tool_choice, BaseModel):
+                    kwargs["tool_choice"] = llm_config.tool_choice.model_dump(
+                        by_alias=True,
+                        exclude_none=True,
+                    )
+                else:
+                    kwargs["tool_choice"] = llm_config.tool_choice
 
         return kwargs
 
diff --git a/sdks/python/oss/tests/pytest/unit/test_prompt_template_extensions.py b/sdks/python/oss/tests/pytest/unit/test_prompt_template_extensions.py
index d29a5246ef..eb620440b0 100644
--- a/sdks/python/oss/tests/pytest/unit/test_prompt_template_extensions.py
+++ b/sdks/python/oss/tests/pytest/unit/test_prompt_template_extensions.py
@@ -1,6 +1,7 @@
 from contextlib import nullcontext
 
 import pytest
+from pydantic import ValidationError
 
 from agenta.sdk.engines.running.handlers import (
     _coerce_fallback_policy,
@@ -182,6 +183,99 @@ def test_null_chat_template_kwargs_is_omitted_from_provider_kwargs():
     assert "chat_template_kwargs" not in prompt.to_openai_kwargs()
 
 
+def test_tool_choice_required_is_passed_through_with_tools():
+    tool = {
+        "type": "function",
+        "function": {
+            "name": "get_weather",
+            "description": "Get weather for a city.",
+            "parameters": {
+                "type": "object",
+                "properties": {"city": {"type": "string"}},
+                "required": ["city"],
+            },
+        },
+    }
+    prompt = PromptTemplate(
+        llm_config=ModelConfig(
+            model="gpt-4o-mini",
+            tools=[tool],
+            tool_choice="required",
+        )
+    )
+
+    kwargs = prompt.to_openai_kwargs()
+
+    assert kwargs["tools"] == [tool]
+    assert kwargs["tool_choice"] == "required"
+
+
+def test_tool_choice_named_function_is_passed_through_with_tools():
+    tool = {
+        "type": "function",
+        "function": {"name": "get_weather", "parameters": {"type": "object"}},
+    }
+    prompt = PromptTemplate(
+        llm_config=ModelConfig(
+            model="gpt-4o-mini",
+            tools=[tool],
+            tool_choice={
+                "type": "function",
+                "function": {"name": "get_weather"},
+            },
+        )
+    )
+
+    assert prompt.to_openai_kwargs()["tool_choice"] == {
+        "type": "function",
+        "function": {"name": "get_weather"},
+    }
+
+
+def test_tool_choice_allowed_tools_is_passed_through_with_tools():
+    allowed_tool = {
+        "type": "function",
+        "function": {"name": "get_weather"},
+    }
+    prompt = PromptTemplate(
+        llm_config=ModelConfig(
+            model="gpt-4o-mini",
+            tools=[allowed_tool],
+            tool_choice={
+                "type": "allowed_tools",
+                "allowed_tools": {
+                    "mode": "required",
+                    "tools": [allowed_tool],
+                },
+            },
+        )
+    )
+
+    assert prompt.to_openai_kwargs()["tool_choice"] == {
+        "type": "allowed_tools",
+        "allowed_tools": {
+            "mode": "required",
+            "tools": [allowed_tool],
+        },
+    }
+
+
+def test_tool_choice_is_omitted_when_tools_are_absent():
+    prompt = PromptTemplate(
+        llm_config=ModelConfig(
+            model="gpt-4o-mini",
+            tool_choice="required",
+        )
+    )
+
+    assert "tool_choice" not in prompt.to_openai_kwargs()
+
+
+def test_tool_choice_rejects_invalid_string_values():
+    with pytest.raises(ValidationError):
+        ModelConfig(model="gpt-4o-mini", tool_choice="always")
+
+
 def test_fallback_config_uses_model_config_defaults():
     prompt = PromptTemplate(fallback_configs=[{"temperature": 0.2}])