feat(vercel_ai_gateway): convert reasoning_effort to Vercel AI Gateway extra_body format

Automatically maps reasoning_effort to extra_body.reasoning = {effort, enabled: true}
as required by Vercel AI Gateway. Adds reasoning_effort to supported params, 2 unit tests,
and provider documentation.

Fixes #25510

Signed-off-by: devarakondasrikanth <devarakondasrikanth@ymail.com>

docs/my-website/docs/providers/vercel_ai_gateway.md

@@ -246,6 +246,36 @@ curl http://localhost:4000/v1/chat/completions \
 
 For more detailed information on using the LiteLLM Proxy, see the [LiteLLM Proxy documentation](../providers/litellm_proxy).
 
+## Reasoning Effort
+
+LiteLLM automatically converts the standard `reasoning_effort` parameter into the format required by Vercel AI Gateway.
+
+**Input (standard LiteLLM format):**
+```python showLineNumbers title="reasoning_effort with Vercel AI Gateway"
+import litellm
+
+response = litellm.completion(
+    model="vercel_ai_gateway/anthropic/claude-sonnet-4-5",
+    messages=[{"role": "user", "content": "Explain quantum entanglement"}],
+    reasoning_effort="high"
+)
+```
+
+LiteLLM converts this to the Vercel AI Gateway `extra_body` format automatically:
+
+```json
+{
+  "extra_body": {
+    "reasoning": {
+      "effort": "high",
+      "enabled": true
+    }
+  }
+}
+```
+
+Supported values for `reasoning_effort`: `"low"`, `"medium"`, `"high"`.
+
 ## Additional Resources
 
 - [Vercel AI Gateway Documentation](https://vercel.com/docs/ai-gateway)

litellm/llms/vercel_ai_gateway/chat/transformation.py

@@ -28,6 +28,8 @@ class VercelAIGatewayConfig(OpenAIGPTConfig):
         base_params = super().get_supported_openai_params(model)
         if "extra_body" not in base_params:
             base_params.append("extra_body")
+        if "reasoning_effort" not in base_params:
+            base_params.append("reasoning_effort")
         return base_params
 
     def _get_openai_compatible_provider_info(
@@ -56,16 +58,20 @@ class VercelAIGatewayConfig(OpenAIGPTConfig):
             non_default_params, optional_params, model, drop_params
         )
 
-        # Vercel AI Gateway-only parameters
-        extra_body = {}
+        # Vercel AI Gateway-only parameters — seed from any extra_body the super call already set
+        extra_body = mapped_openai_params.pop("extra_body", {})
         provider_options = non_default_params.pop("providerOptions", None)
 
         if provider_options is not None:
             extra_body["providerOptions"] = provider_options
 
-        mapped_openai_params[
-            "extra_body"
-        ] = extra_body  # openai client supports `extra_body` param
+        reasoning_effort = mapped_openai_params.pop("reasoning_effort", None)
+        if reasoning_effort is not None:
+            extra_body["reasoning"] = {"effort": reasoning_effort, "enabled": True}
+
+        mapped_openai_params["extra_body"] = (
+            extra_body  # openai client supports `extra_body` param
+        )
         return mapped_openai_params
 
     def transform_request(

tests/test_litellm/llms/vercel_ai_gateway/chat/test_vercel_ai_gateway_transformation.py

@@ -2,8 +2,6 @@ import os
 import sys
 from unittest.mock import patch
 
-import pytest
-
 sys.path.insert(
     0, os.path.abspath("../../../../..")
 )  # Adds the parent directory to the system path
@@ -21,16 +19,17 @@ def test_vercel_ai_gateway_extra_body_transformation():
         messages=[{"role": "user", "content": "Hello, world!"}],
         optional_params={
             "extra_body": {
-                "providerOptions": {
-                    "gateway": {"order": ["azure", "openai"]}
-                }
+                "providerOptions": {"gateway": {"order": ["azure", "openai"]}}
             }
         },
         litellm_params={},
         headers={},
     )
 
-    assert transformed_request["extra_body"]["providerOptions"]["gateway"]["order"] == ["azure", "openai"]
+    assert transformed_request["extra_body"]["providerOptions"]["gateway"]["order"] == [
+        "azure",
+        "openai",
+    ]
     assert transformed_request["messages"] == [
         {"role": "user", "content": "Hello, world!"}
     ]
@@ -39,28 +38,31 @@ def test_vercel_ai_gateway_extra_body_transformation():
 def test_vercel_ai_gateway_provider_options_mapping():
     """Test that providerOptions from non_default_params is moved to extra_body"""
     config = VercelAIGatewayConfig()
-    
+
     non_default_params = {
-        "providerOptions": {
-            "gateway": {"order": ["azure", "openai"]}
-        }
+        "providerOptions": {"gateway": {"order": ["azure", "openai"]}}
     }
     optional_params = {}
     model = "vercel_ai_gateway/openai/gpt-4o"
-    
+
     result = config.map_openai_params(
         non_default_params, optional_params, model, drop_params=False
     )
-    
-    assert result["extra_body"]["providerOptions"]["gateway"]["order"] == ["azure", "openai"]
+
+    assert result["extra_body"]["providerOptions"]["gateway"]["order"] == [
+        "azure",
+        "openai",
+    ]
     assert "providerOptions" not in result
 
 
 def test_vercel_ai_gateway_get_supported_openai_params():
     """Test that extra_body is included in supported params"""
     config = VercelAIGatewayConfig()
-    supported_params = config.get_supported_openai_params("vercel_ai_gateway/openai/gpt-4o")
-    
+    supported_params = config.get_supported_openai_params(
+        "vercel_ai_gateway/openai/gpt-4o"
+    )
+
     assert "extra_body" in supported_params
     assert "temperature" in supported_params
     assert "max_tokens" in supported_params
@@ -70,7 +72,7 @@ def test_vercel_ai_gateway_get_supported_openai_params():
 def test_vercel_ai_gateway_get_openai_compatible_provider_info():
     """Test provider info retrieval with environment variables"""
     config = VercelAIGatewayConfig()
-    
+
     with patch.dict(
         "os.environ",
         {
@@ -86,27 +88,72 @@ def test_vercel_ai_gateway_get_openai_compatible_provider_info():
 def test_vercel_ai_gateway_error_class():
     """Test error class creation"""
     config = VercelAIGatewayConfig()
-    
+
     error_message = "Test error"
     status_code = 400
     headers = {"Content-Type": "application/json"}
-    
+
     error_class = config.get_error_class(error_message, status_code, headers)
-    
+
     assert isinstance(error_class, VercelAIGatewayException)
     assert error_class.message == error_message
     assert error_class.status_code == status_code
     assert error_class.headers == headers
 
 
+def test_vercel_ai_gateway_reasoning_effort_mapping():
+    """Test that reasoning_effort is converted to extra_body.reasoning format"""
+    config = VercelAIGatewayConfig()
+
+    non_default_params = {"reasoning_effort": "high"}
+    optional_params = {}
+    model = "vercel_ai_gateway/openai/gpt-4o"
+
+    result = config.map_openai_params(
+        non_default_params, optional_params, model, drop_params=False
+    )
+
+    assert result["extra_body"]["reasoning"] == {"effort": "high", "enabled": True}
+    assert "reasoning_effort" not in result
+
+
+def test_vercel_ai_gateway_reasoning_effort_with_existing_extra_body():
+    """Test that reasoning_effort merges with existing extra_body instead of overwriting it"""
+    config = VercelAIGatewayConfig()
+
+    non_default_params = {
+        "reasoning_effort": "high",
+        "providerOptions": {"gateway": {"order": ["azure", "openai"]}},
+    }
+    optional_params = {}
+    model = "vercel_ai_gateway/openai/gpt-4o"
+
+    result = config.map_openai_params(
+        non_default_params, optional_params, model, drop_params=False
+    )
+
+    # Both keys must coexist in extra_body — neither should be silently dropped
+    assert result["extra_body"]["reasoning"] == {"effort": "high", "enabled": True}
+    assert result["extra_body"]["providerOptions"] == {
+        "gateway": {"order": ["azure", "openai"]}
+    }
+    assert "reasoning_effort" not in result
+
+
+def test_vercel_ai_gateway_reasoning_effort_in_supported_params():
+    """Test that reasoning_effort is included in supported params"""
+    config = VercelAIGatewayConfig()
+    supported_params = config.get_supported_openai_params(
+        "vercel_ai_gateway/openai/gpt-4o"
+    )
+
+    assert "reasoning_effort" in supported_params
+
+
 def test_vercel_ai_gateway_exception_inheritance():
     """Test that VercelAIGatewayException inherits from BaseLLMException"""
     from litellm.llms.base_llm.chat.transformation import BaseLLMException
-    
-    exception = VercelAIGatewayException(
-        message="test", 
-        status_code=500, 
-        headers={}
-    )
-    
+
+    exception = VercelAIGatewayException(message="test", status_code=500, headers={})
+
     assert isinstance(exception, BaseLLMException)