diff --git a/docs/features/tool_calling.md b/docs/features/tool_calling.md
index cea117541..e9aa87a69 100644
--- a/docs/features/tool_calling.md
+++ b/docs/features/tool_calling.md
@@ -505,7 +505,7 @@ Here is a summary of a plugin file:
 
         # adjust request. e.g.: set skip special tokens
         # to False for tool call output.
-        def adjust_request(self, request: ChatCompletionRequest) -> ChatCompletionRequest:
+        def adjust_request(self, request: ChatCompletionRequest | ResponsesRequest) -> ChatCompletionRequest | ResponsesRequest:
             return request
 
         # implement the tool call parse for stream call
diff --git a/vllm/entrypoints/serve/render/serving.py b/vllm/entrypoints/serve/render/serving.py
index 52f03447d..83b41bbda 100644
--- a/vllm/entrypoints/serve/render/serving.py
+++ b/vllm/entrypoints/serve/render/serving.py
@@ -546,7 +546,7 @@ class OpenAIServingRender:
                     raise NotImplementedError(msg)
                 tokenizer = renderer.get_tokenizer()
                 request = tool_parser(tokenizer, request.tools).adjust_request(
-                    request=request  # type: ignore[arg-type]
+                    request=request
                 )
 
         return conversation, [engine_input]
diff --git a/vllm/parser/abstract_parser.py b/vllm/parser/abstract_parser.py
index dd9dc9423..8a12f6fe1 100644
--- a/vllm/parser/abstract_parser.py
+++ b/vllm/parser/abstract_parser.py
@@ -32,9 +32,7 @@ from vllm.entrypoints.openai.engine.protocol import (
     FunctionCall,
     FunctionDefinition,
 )
-from vllm.entrypoints.openai.responses.protocol import (
-    ResponsesRequest,
-)
+from vllm.entrypoints.openai.responses.protocol import ResponsesRequest
 from vllm.logger import init_logger
 from vllm.reasoning.abs_reasoning_parsers import ReasoningParser
 from vllm.tokenizers import TokenizerLike
@@ -229,7 +227,9 @@ class Parser:
 
     # ========== Tool Parser Methods ==========
 
-    def adjust_request(self, request: ChatCompletionRequest) -> ChatCompletionRequest:
+    def adjust_request(
+        self, request: ChatCompletionRequest | ResponsesRequest
+    ) -> ChatCompletionRequest | ResponsesRequest:
         """
         Adjust the request parameters for tool calling.
 
diff --git a/vllm/tool_parsers/abstract_tool_parser.py b/vllm/tool_parsers/abstract_tool_parser.py
index 3f578b1b6..9c05bb6b9 100644
--- a/vllm/tool_parsers/abstract_tool_parser.py
+++ b/vllm/tool_parsers/abstract_tool_parser.py
@@ -62,7 +62,9 @@ class ToolParser:
         # whereas all tokenizers have .get_vocab()
         return self.model_tokenizer.get_vocab()
 
-    def adjust_request(self, request: ChatCompletionRequest) -> ChatCompletionRequest:
+    def adjust_request(
+        self, request: ChatCompletionRequest | ResponsesRequest
+    ) -> ChatCompletionRequest | ResponsesRequest:
         """
         Static method that used to adjust the request parameters.
         """
diff --git a/vllm/tool_parsers/deepseekv32_tool_parser.py b/vllm/tool_parsers/deepseekv32_tool_parser.py
index a9772b753..f30231009 100644
--- a/vllm/tool_parsers/deepseekv32_tool_parser.py
+++ b/vllm/tool_parsers/deepseekv32_tool_parser.py
@@ -19,6 +19,7 @@ from vllm.entrypoints.openai.engine.protocol import (
     FunctionCall,
     ToolCall,
 )
+from vllm.entrypoints.openai.responses.protocol import ResponsesRequest
 from vllm.logger import init_logger
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.abstract_tool_parser import (
@@ -78,7 +79,9 @@ class DeepSeekV32ToolParser(ToolParser):
             "vLLM Successfully import tool parser %s !", self.__class__.__name__
         )
 
-    def adjust_request(self, request):
+    def adjust_request(
+        self, request: ChatCompletionRequest | ResponsesRequest
+    ) -> ChatCompletionRequest | ResponsesRequest:
         request = super().adjust_request(request)
         if request.tools and request.tool_choice != "none":
             # Ensure tool call tokens
diff --git a/vllm/tool_parsers/functiongemma_tool_parser.py b/vllm/tool_parsers/functiongemma_tool_parser.py
index dfd91d974..35c4c6b84 100644
--- a/vllm/tool_parsers/functiongemma_tool_parser.py
+++ b/vllm/tool_parsers/functiongemma_tool_parser.py
@@ -18,6 +18,7 @@ from vllm.entrypoints.openai.engine.protocol import (
     FunctionCall,
     ToolCall,
 )
+from vllm.entrypoints.openai.responses.protocol import ResponsesRequest
 from vllm.logger import init_logger
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.abstract_tool_parser import Tool, ToolParser
@@ -86,7 +87,9 @@ class FunctionGemmaToolParser(ToolParser):
 
         return arguments
 
-    def adjust_request(self, request: ChatCompletionRequest) -> ChatCompletionRequest:
+    def adjust_request(
+        self, request: ChatCompletionRequest | ResponsesRequest
+    ) -> ChatCompletionRequest | ResponsesRequest:
         request = super().adjust_request(request)
         if request.tools and request.tool_choice != "none":
             request.skip_special_tokens = False
diff --git a/vllm/tool_parsers/gigachat3_tool_parser.py b/vllm/tool_parsers/gigachat3_tool_parser.py
index f470f6a5b..d48ff43ee 100644
--- a/vllm/tool_parsers/gigachat3_tool_parser.py
+++ b/vllm/tool_parsers/gigachat3_tool_parser.py
@@ -18,6 +18,7 @@ from vllm.entrypoints.openai.engine.protocol import (
     FunctionCall,
     ToolCall,
 )
+from vllm.entrypoints.openai.responses.protocol import ResponsesRequest
 from vllm.logger import init_logger
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.abstract_tool_parser import Tool, ToolParser
@@ -55,7 +56,9 @@ class GigaChat3ToolParser(ToolParser):
         self.end_content: bool = False
         self.streamed_args_for_tool: list[str] = []
 
-    def adjust_request(self, request: ChatCompletionRequest) -> ChatCompletionRequest:
+    def adjust_request(
+        self, request: ChatCompletionRequest | ResponsesRequest
+    ) -> ChatCompletionRequest | ResponsesRequest:
         request = super().adjust_request(request)
         if request.tools and request.tool_choice != "none":
             request.skip_special_tokens = False
diff --git a/vllm/tool_parsers/glm4_moe_tool_parser.py b/vllm/tool_parsers/glm4_moe_tool_parser.py
index fc718921d..d1778c565 100644
--- a/vllm/tool_parsers/glm4_moe_tool_parser.py
+++ b/vllm/tool_parsers/glm4_moe_tool_parser.py
@@ -30,6 +30,7 @@ from vllm.entrypoints.openai.engine.protocol import (
     FunctionCall,
     ToolCall,
 )
+from vllm.entrypoints.openai.responses.protocol import ResponsesRequest
 from vllm.logger import init_logger
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.abstract_tool_parser import (
@@ -151,7 +152,9 @@ class Glm4MoeModelToolParser(ToolParser):
             logger.exception("Failed to determine if tools are enabled.")
             return False
 
-    def adjust_request(self, request: ChatCompletionRequest) -> ChatCompletionRequest:
+    def adjust_request(
+        self, request: ChatCompletionRequest | ResponsesRequest
+    ) -> ChatCompletionRequest | ResponsesRequest:
         """Adjust request parameters for tool call token handling."""
         request = super().adjust_request(request)
         if request.tools and request.tool_choice != "none":
diff --git a/vllm/tool_parsers/granite4_tool_parser.py b/vllm/tool_parsers/granite4_tool_parser.py
index 3d58690f5..b11e9e6d5 100644
--- a/vllm/tool_parsers/granite4_tool_parser.py
+++ b/vllm/tool_parsers/granite4_tool_parser.py
@@ -19,6 +19,7 @@ from vllm.entrypoints.openai.engine.protocol import (
     FunctionCall,
     ToolCall,
 )
+from vllm.entrypoints.openai.responses.protocol import ResponsesRequest
 from vllm.logger import init_logger
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.abstract_tool_parser import (
@@ -59,7 +60,9 @@ class Granite4ToolParser(ToolParser):
         self.start_regex = re.compile(self.tc_start)
         self.end_regex = re.compile(self.tc_end)
 
-    def adjust_request(self, request: ChatCompletionRequest) -> ChatCompletionRequest:
+    def adjust_request(
+        self, request: ChatCompletionRequest | ResponsesRequest
+    ) -> ChatCompletionRequest | ResponsesRequest:
         request = super().adjust_request(request)
         if request.tools and request.tool_choice != "none":
             # do not skip special tokens because the tool_call tokens are
diff --git a/vllm/tool_parsers/hermes_tool_parser.py b/vllm/tool_parsers/hermes_tool_parser.py
index 4e54d75b4..469bc8363 100644
--- a/vllm/tool_parsers/hermes_tool_parser.py
+++ b/vllm/tool_parsers/hermes_tool_parser.py
@@ -18,6 +18,7 @@ from vllm.entrypoints.openai.engine.protocol import (
     FunctionCall,
     ToolCall,
 )
+from vllm.entrypoints.openai.responses.protocol import ResponsesRequest
 from vllm.logger import init_logger
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.abstract_tool_parser import (
@@ -77,7 +78,9 @@ class Hermes2ProToolParser(ToolParser):
         # Streaming state: what has been sent to the client.
         self._sent_content_idx: int = 0
 
-    def adjust_request(self, request: ChatCompletionRequest) -> ChatCompletionRequest:
+    def adjust_request(
+        self, request: ChatCompletionRequest | ResponsesRequest
+    ) -> ChatCompletionRequest | ResponsesRequest:
         request = super().adjust_request(request)
         if request.tools and request.tool_choice != "none":
             # do not skip special tokens because the tool_call tokens are
diff --git a/vllm/tool_parsers/internlm2_tool_parser.py b/vllm/tool_parsers/internlm2_tool_parser.py
index fc7c44cff..6f38b851a 100644
--- a/vllm/tool_parsers/internlm2_tool_parser.py
+++ b/vllm/tool_parsers/internlm2_tool_parser.py
@@ -19,6 +19,7 @@ from vllm.entrypoints.openai.engine.protocol import (
     FunctionCall,
     ToolCall,
 )
+from vllm.entrypoints.openai.responses.protocol import ResponsesRequest
 from vllm.logger import init_logger
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.abstract_tool_parser import (
@@ -35,7 +36,9 @@ class Internlm2ToolParser(ToolParser):
         super().__init__(tokenizer, tools)
         self.position = 0
 
-    def adjust_request(self, request: ChatCompletionRequest) -> ChatCompletionRequest:
+    def adjust_request(
+        self, request: ChatCompletionRequest | ResponsesRequest
+    ) -> ChatCompletionRequest | ResponsesRequest:
         request = super().adjust_request(request)
         if request.tools and request.tool_choice != "none":
             # do not skip special tokens because internlm use the special
diff --git a/vllm/tool_parsers/jamba_tool_parser.py b/vllm/tool_parsers/jamba_tool_parser.py
index 5a9af9910..dec3c88d9 100644
--- a/vllm/tool_parsers/jamba_tool_parser.py
+++ b/vllm/tool_parsers/jamba_tool_parser.py
@@ -20,6 +20,7 @@ from vllm.entrypoints.openai.engine.protocol import (
     FunctionCall,
     ToolCall,
 )
+from vllm.entrypoints.openai.responses.protocol import ResponsesRequest
 from vllm.logger import init_logger
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.abstract_tool_parser import Tool, ToolParser
@@ -68,7 +69,9 @@ class JambaToolParser(ToolParser):
                 "tokens in the tokenizer!"
             )
 
-    def adjust_request(self, request: ChatCompletionRequest) -> ChatCompletionRequest:
+    def adjust_request(
+        self, request: ChatCompletionRequest | ResponsesRequest
+    ) -> ChatCompletionRequest | ResponsesRequest:
         request = super().adjust_request(request)
         if request.tools and request.tool_choice != "none":
             # do not skip special tokens because jamba use the special
diff --git a/vllm/tool_parsers/mistral_tool_parser.py b/vllm/tool_parsers/mistral_tool_parser.py
index 153c6ed32..dc92522a0 100644
--- a/vllm/tool_parsers/mistral_tool_parser.py
+++ b/vllm/tool_parsers/mistral_tool_parser.py
@@ -23,6 +23,7 @@ from vllm.entrypoints.openai.engine.protocol import (
     FunctionCall,
     ToolCall,
 )
+from vllm.entrypoints.openai.responses.protocol import ResponsesRequest
 from vllm.logger import init_logger
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.abstract_tool_parser import (
@@ -111,7 +112,9 @@ class MistralToolParser(ToolParser):
                 "the tokenizer!"
             )
 
-    def adjust_request(self, request: ChatCompletionRequest) -> ChatCompletionRequest:
+    def adjust_request(
+        self, request: ChatCompletionRequest | ResponsesRequest
+    ) -> ChatCompletionRequest | ResponsesRequest:
         request = super().adjust_request(request)
         if (
             not is_mistral_tokenizer(self.model_tokenizer)
diff --git a/vllm/tool_parsers/step3_tool_parser.py b/vllm/tool_parsers/step3_tool_parser.py
index a9c569587..06e7a6466 100644
--- a/vllm/tool_parsers/step3_tool_parser.py
+++ b/vllm/tool_parsers/step3_tool_parser.py
@@ -19,6 +19,7 @@ from vllm.entrypoints.openai.engine.protocol import (
     FunctionCall,
     ToolCall,
 )
+from vllm.entrypoints.openai.responses.protocol import ResponsesRequest
 from vllm.logger import init_logger
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.abstract_tool_parser import (
@@ -51,7 +52,9 @@ class Step3ToolParser(ToolParser):
         self.tool_block_started = False
         self.tool_block_finished = False
 
-    def adjust_request(self, request: ChatCompletionRequest) -> ChatCompletionRequest:
+    def adjust_request(
+        self, request: ChatCompletionRequest | ResponsesRequest
+    ) -> ChatCompletionRequest | ResponsesRequest:
         request = super().adjust_request(request)
         if request.tools and request.tool_choice != "none":
             request.skip_special_tokens = False