diff --git a/cecli/coders/base_coder.py b/cecli/coders/base_coder.py
index e7ef8ebec20..0a11adecdd4 100755
--- a/cecli/coders/base_coder.py
+++ b/cecli/coders/base_coder.py
@@ -598,6 +598,9 @@ def __init__(
                 self.io.tool_output("JSON Schema:")
                 self.io.tool_output(json.dumps(self.functions, indent=4))
 
+        # Track partial response state for each request
+        self._reset_partial_response_flags()
+
     @property
     def gpt_prompts(self):
         """Get prompts from the registry based on the coder type."""
@@ -3013,6 +3016,7 @@ async def send(self, messages, model=None, functions=None, tools=None):
         self.partial_response_chunks = []
         self.partial_response_tool_calls = []
         self.partial_response_function_call = dict()
+        self._reset_partial_response_flags()
 
         completion = None
         self.token_profiler.start()
@@ -3058,6 +3062,84 @@ async def send(self, messages, model=None, functions=None, tools=None):
                 if args:
                     self.io.ai_output(json.dumps(args, indent=4))
 
+    def _reset_partial_response_flags(self):
+        self._partial_response_received_flags = {
+            "content": False,
+            "reasoning": False,
+            "tool_calls": False,
+            "function_call": False,
+        }
+
+    def _register_partial_response(
+        self,
+        *,
+        content=False,
+        reasoning=False,
+        tool_calls=False,
+        function_call=False,
+    ):
+        if not hasattr(self, "_partial_response_received_flags"):
+            self._reset_partial_response_flags()
+
+        flags = self._partial_response_received_flags
+        if content:
+            flags["content"] = True
+        if reasoning:
+            flags["reasoning"] = True
+        if tool_calls:
+            flags["tool_calls"] = True
+        if function_call:
+            flags["function_call"] = True
+
+    def _received_any_partial_response(self, received_content_flag=False):
+        if not hasattr(self, "_partial_response_received_flags"):
+            return False
+
+        flags = self._partial_response_received_flags
+        if received_content_flag:
+            return True
+        return any(flags.values())
+
+    def _get_empty_response_message(self):
+        """Generate a descriptive warning for empty responses."""
+        flags = getattr(self, "_partial_response_received_flags", {})
+
+        has_content = flags.get("content", False)
+        has_reasoning = flags.get("reasoning", False)
+        has_tool_calls = flags.get("tool_calls", False)
+        has_function_call = flags.get("function_call", False)
+
+        if has_tool_calls and not has_content:
+            if has_reasoning:
+                return (
+                    "Empty response received from LLM. "
+                    "Only tool calls and reasoning content were received, but no text response. "
+                    "Check if the model is configured to return text content."
+                )
+            return (
+                "Empty response received from LLM. "
+                "Only tool calls were received, but no text response. "
+                "Check if the model is configured to return text content."
+            )
+        if has_reasoning and not has_content:
+            return (
+                "Empty response received from LLM. "
+                "Only reasoning content was received, but no text response. "
+                "Check if the model is configured to return text content."
+            )
+        if has_function_call and not has_content:
+            return (
+                "Empty response received from LLM. "
+                "Only function calls were received, but no text response. "
+                "Check if the model is configured to return text content."
+            )
+
+        return (
+            "Empty response received from LLM. "
+            "No content, tool calls, or reasoning was received. "
+            "Check your provider account, model availability, or network connectivity."
+        )
+
     def show_send_output(self, completion):
         if self.verbose:
             print(completion)
@@ -3072,6 +3154,31 @@ def show_send_output(self, completion):
 
         self.partial_response_chunks.append(completion)
 
+        try:
+            message = None
+            if completion.choices and completion.choices[0].message:
+                message = completion.choices[0].message
+        except (AttributeError, IndexError):
+            message = None
+
+        if message:
+            if getattr(message, "tool_calls", None):
+                self._register_partial_response(tool_calls=True)
+
+            if getattr(message, "function_call", None):
+                self._register_partial_response(function_call=True)
+
+            reasoning_content = getattr(message, "reasoning_content", None)
+            if reasoning_content:
+                self._register_partial_response(reasoning=True)
+            else:
+                reasoning_attr = getattr(message, "reasoning", None)
+                if reasoning_attr:
+                    self._register_partial_response(reasoning=True)
+
+            if getattr(message, "content", None):
+                self._register_partial_response(content=True)
+
         response, func_err, content_err = self.consolidate_chunks()
 
         resp_hash = dict(
@@ -3096,7 +3203,10 @@ def show_send_output(self, completion):
 
         show_resp = replace_reasoning_tags(show_resp, self.reasoning_tag_name)
 
-        self.io.assistant_output(show_resp, pretty=self.show_pretty())
+        if show_resp:
+            self.io.assistant_output(show_resp, pretty=self.show_pretty())
+        elif not self._received_any_partial_response():
+            self.io.tool_warning(self._get_empty_response_message())
 
         if (
             hasattr(completion.choices[0], "finish_reason")
@@ -3243,6 +3353,8 @@ def consolidate_chunks(self):
         for chunk in self.partial_response_chunks:
             try:
                 if chunk.choices and chunk.choices[0].delta and chunk.choices[0].delta.tool_calls:
+                    if not self.stream:
+                        self._register_partial_response(tool_calls=True)
                     for tool_call in chunk.choices[0].delta.tool_calls:
                         if (
                             hasattr(tool_call, "provider_specific_fields")
@@ -3271,6 +3383,8 @@ def consolidate_chunks(self):
 
         try:
             if response.choices[0].message.tool_calls:
+                if not self.stream:
+                    self._register_partial_response(tool_calls=True)
                 for i, tool_call in enumerate(response.choices[0].message.tool_calls):
                     # Add provider-specific fields if we collected any for this tool
                     tool_id = tool_call.id
@@ -3304,6 +3418,8 @@ def consolidate_chunks(self):
                 self.partial_response_function_call = (
                     response.choices[0].message.tool_calls[0].function
                 )
+                if self.partial_response_function_call and not self.stream:
+                    self._register_partial_response(function_call=True)
         except AttributeError as e:
             func_err = e
 
@@ -3315,7 +3431,12 @@ def consolidate_chunks(self):
             except AttributeError:
                 reasoning_content = None
 
-        self.partial_response_reasoning_content = reasoning_content or ""
+        if reasoning_content:
+            self.partial_response_reasoning_content = reasoning_content
+            if not self.stream:
+                self._register_partial_response(reasoning=True)
+        else:
+            self.partial_response_reasoning_content = ""
 
         try:
             content = response.choices[0].message.content
@@ -3331,7 +3452,12 @@ def consolidate_chunks(self):
                     for block in content
                     if isinstance(block, dict) and block.get("type") == "text"
                 )
-            self.partial_response_content = content or ""
+            if content:
+                self.partial_response_content = content
+                if not self.stream:
+                    self._register_partial_response(content=True)
+            else:
+                self.partial_response_content = ""
         except AttributeError as e:
             content_err = e
 
diff --git a/cecli/io.py b/cecli/io.py
index 482eebccf95..d7544c64faa 100644
--- a/cecli/io.py
+++ b/cecli/io.py
@@ -1484,7 +1484,10 @@ def profile(self, *messages, start=False):
 
     def assistant_output(self, message, pretty=None):
         if not message:
-            self.tool_warning("Empty response received from LLM. Check your provider account?")
+            self.tool_warning(
+                "Empty response received from LLM. No text content was returned. "
+                "Check your provider account, model availability, or network connectivity."
+            )
             return
 
         show_resp = message
diff --git a/cecli/tui/io.py b/cecli/tui/io.py
index 6ae017eaeb6..3a58326e4ed 100644
--- a/cecli/tui/io.py
+++ b/cecli/tui/io.py
@@ -170,7 +170,10 @@ def assistant_output(self, message, pretty=None):
             pretty: Whether to use pretty formatting (unused in TUI, kept for compatibility)
         """
         if not message:
-            self.tool_warning("Empty response received from LLM. Check your provider account?")
+            self.tool_warning(
+                "Empty response received from LLM. No text content was returned. "
+                "Check your provider account, model availability, or network connectivity."
+            )
             return
 
         # Use the streaming path so markdown rendering is applied
diff --git a/tests/basic/test_coder.py b/tests/basic/test_coder.py
index febe38028e3..48fa9f613ff 100644
--- a/tests/basic/test_coder.py
+++ b/tests/basic/test_coder.py
@@ -1927,3 +1927,118 @@ async def test_execute_tool_calls_blob_content(self, mock_call_openai_tool):
                 " (application/octet-stream)]"
             )
             assert result[0]["content"] == expected_content
+
+    async def test_reset_partial_response_flags(self):
+        """Test that _reset_partial_response_flags initializes all flags to False."""
+        with GitTemporaryDirectory():
+            io = InputOutput(yes=True)
+            coder = await Coder.create(self.GPT35, "diff", io=io)
+
+            coder._reset_partial_response_flags()
+            flags = coder._partial_response_received_flags
+            assert flags["content"] is False
+            assert flags["reasoning"] is False
+            assert flags["tool_calls"] is False
+            assert flags["function_call"] is False
+
+    async def test_register_partial_response(self):
+        """Test that _register_partial_response correctly sets flags."""
+        with GitTemporaryDirectory():
+            io = InputOutput(yes=True)
+            coder = await Coder.create(self.GPT35, "diff", io=io)
+
+            coder._reset_partial_response_flags()
+            coder._register_partial_response(content=True)
+            assert coder._partial_response_received_flags["content"] is True
+            assert coder._partial_response_received_flags["reasoning"] is False
+
+            coder._register_partial_response(reasoning=True)
+            assert coder._partial_response_received_flags["reasoning"] is True
+
+            coder._register_partial_response(tool_calls=True)
+            assert coder._partial_response_received_flags["tool_calls"] is True
+
+            coder._register_partial_response(function_call=True)
+            assert coder._partial_response_received_flags["function_call"] is True
+
+            coder._reset_partial_response_flags()
+            coder._register_partial_response(content=True, reasoning=True)
+            assert coder._partial_response_received_flags["content"] is True
+            assert coder._partial_response_received_flags["reasoning"] is True
+            assert coder._partial_response_received_flags["tool_calls"] is False
+
+    async def test_received_any_partial_response(self):
+        """Test that _received_any_partial_response correctly checks flags."""
+        with GitTemporaryDirectory():
+            io = InputOutput(yes=True)
+            coder = await Coder.create(self.GPT35, "diff", io=io)
+
+            coder._reset_partial_response_flags()
+            assert coder._received_any_partial_response() is False
+            assert coder._received_any_partial_response(received_content_flag=True) is True
+
+            coder._register_partial_response(content=True)
+            assert coder._received_any_partial_response() is True
+
+            coder._reset_partial_response_flags()
+            coder._register_partial_response(reasoning=True)
+            assert coder._received_any_partial_response() is True
+
+            coder._reset_partial_response_flags()
+            coder._register_partial_response(tool_calls=True)
+            assert coder._received_any_partial_response() is True
+
+            coder._reset_partial_response_flags()
+            coder._register_partial_response(function_call=True)
+            assert coder._received_any_partial_response() is True
+
+    async def test_get_empty_response_message_variants(self):
+        """Verify _get_empty_response_message returns descriptive strings."""
+        with GitTemporaryDirectory():
+            io = InputOutput(yes=True)
+            coder = await Coder.create(self.GPT35, "diff", io=io)
+
+            coder._reset_partial_response_flags()
+            coder._register_partial_response(tool_calls=True)
+            msg = coder._get_empty_response_message()
+            assert "Only tool calls" in msg
+
+            coder._reset_partial_response_flags()
+            coder._register_partial_response(tool_calls=True, reasoning=True)
+            msg = coder._get_empty_response_message()
+            assert "tool calls and reasoning" in msg
+
+            coder._reset_partial_response_flags()
+            coder._register_partial_response(reasoning=True)
+            msg = coder._get_empty_response_message()
+            assert "Only reasoning content" in msg
+
+            coder._reset_partial_response_flags()
+            coder._register_partial_response(function_call=True)
+            msg = coder._get_empty_response_message()
+            assert "Only function calls" in msg
+
+    async def test_initialization_in_constructor(self):
+        """Test that coders initialize the partial response flags."""
+        with GitTemporaryDirectory():
+            io = InputOutput(yes=True)
+            coder = await Coder.create(self.GPT35, "diff", io=io)
+
+            assert "_partial_response_received_flags" in coder.__dict__
+            flags = coder._partial_response_received_flags
+            assert flags["content"] is False
+            assert flags["reasoning"] is False
+            assert flags["tool_calls"] is False
+            assert flags["function_call"] is False
+
+    async def test_received_content_flag_override(self):
+        """Test that received_content_flag overrides other flags."""
+        with GitTemporaryDirectory():
+            io = InputOutput(yes=True)
+            coder = await Coder.create(self.GPT35, "diff", io=io)
+
+            coder._reset_partial_response_flags()
+            assert coder._received_any_partial_response(received_content_flag=True) is True
+
+            coder._register_partial_response(content=True, reasoning=True)
+            assert coder._received_any_partial_response(received_content_flag=True) is True