diff --git a/ui/litellm-dashboard/src/app/chat/page.integration.test.tsx b/ui/litellm-dashboard/src/app/chat/page.integration.test.tsx
index c0b6de2df9e..16a9be54292 100644
--- a/ui/litellm-dashboard/src/app/chat/page.integration.test.tsx
+++ b/ui/litellm-dashboard/src/app/chat/page.integration.test.tsx
@@ -1,4 +1,4 @@
-import { fireEvent, screen, waitFor, within } from "@testing-library/react";
+import { act, fireEvent, screen, waitFor, within } from "@testing-library/react";
 import { beforeEach, describe, expect, it, vi } from "vitest";
 import { useChatHistory } from "@/components/chat/useChatHistory";
 import ChatConversationPage from "./page";
@@ -154,6 +154,44 @@ describe("/ui/chat request metrics", () => {
     ]);
   });
 
+  it.each([false, true])("ignores late response IDs after model changes ($0)", async (switchBack) => {
+    vi.mocked(fetchAvailableModels).mockResolvedValue([
+      { model_group: "gpt-5.4-mini", providers: ["openai"] },
+      { model_group: "second-model", providers: ["openai"] },
+    ]);
+    const firstCompletion = Promise.withResolvers<void>();
+    mockMakeOpenAIResponsesRequest.mockImplementationOnce(async (...args: unknown[]) => {
+      await firstCompletion.promise;
+      (args[1] as (role: string, delta: string) => void)("assistant", "First answer");
+      (args[15] as (id: string) => void)("response-first");
+    });
+    mockMakeOpenAIResponsesRequest.mockResolvedValue(undefined);
+
+    await sendOneMessage();
+    fireEvent.click(screen.getByRole("button", { name: /gpt-5\.4-mini/ }));
+    fireEvent.click(await screen.findByRole("button", { name: "second-model" }));
+    if (switchBack) {
+      fireEvent.click(screen.getByRole("button", { name: "second-model" }));
+      fireEvent.click(await screen.findByRole("button", { name: "gpt-5.4-mini" }));
+    }
+
+    await act(async () => firstCompletion.resolve());
+    expect(await screen.findByText("First answer")).toBeVisible();
+    fireEvent.change(screen.getByRole("textbox"), { target: { value: "Follow up" } });
+    await waitFor(() => expect(screen.getByRole("button", { name: "Send" })).toBeEnabled());
+    fireEvent.click(screen.getByRole("button", { name: "Send" }));
+    await waitFor(() => expect(mockMakeOpenAIResponsesRequest).toHaveBeenCalledTimes(2));
+
+    const second = mockMakeOpenAIResponsesRequest.mock.calls[1];
+    expect(second[2]).toBe(switchBack ? "gpt-5.4-mini" : "second-model");
+    expect(second[14]).toBeNull();
+    expect(second[0]).toEqual([
+      { role: "user", content: "How much did this cost?" },
+      { role: "assistant", content: "First answer" },
+      { role: "user", content: "Follow up" },
+    ]);
+  });
+
   it("renders latency, TTFT, token counts and cost reported for the assistant turn", async () => {
     mockMakeOpenAIResponsesRequest.mockImplementation(async (...args: unknown[]) => {
       const updateTextUI = args[1] as (role: string, delta: string) => void;
diff --git a/ui/litellm-dashboard/src/app/chat/page.tsx b/ui/litellm-dashboard/src/app/chat/page.tsx
index 256031c2c41..def1223ec0c 100644
--- a/ui/litellm-dashboard/src/app/chat/page.tsx
+++ b/ui/litellm-dashboard/src/app/chat/page.tsx
@@ -77,6 +77,7 @@ export default function ChatConversationPage() {
   const [modelSearchText, setModelSearchText] = useState("");
 
   const [responsesSessionId, setResponsesSessionId] = useState<string | null>(null);
+  const responsesSessionGenerationRef = useRef(0);
   const [prevConversationIdForSessionReset, setPrevConversationIdForSessionReset] = useState(activeConversationId);
   const [isStreaming, setIsStreaming] = useState(false);
   const [inputText, setInputText] = useState("");
@@ -129,6 +130,7 @@ export default function ChatConversationPage() {
   }
 
   const selectModel = useCallback((model: string) => {
+    responsesSessionGenerationRef.current += 1;
     setSelectedModel(model);
     setResponsesSessionId(null);
     localStorage.setItem(LOCALSTORAGE_MODEL_KEY, model);
@@ -141,6 +143,7 @@ export default function ChatConversationPage() {
       const trimmed = text.trim();
       if (!trimmed || !selectedModel || isStreaming) return;
       const model = selectedModel;
+      const sessionGeneration = responsesSessionGenerationRef.current;
       setInputText("");
 
       let convId = activeConversationId;
@@ -213,7 +216,11 @@ export default function ChatConversationPage() {
           undefined,
           selectedMCPServers.length > 0 ? selectedMCPServers : undefined,
           previousResponseId,
-          (id: string) => setResponsesSessionId(id),
+          (id: string) => {
+            if (responsesSessionGenerationRef.current === sessionGeneration) {
+              setResponsesSessionId(id);
+            }
+          },
           (event: MCPEvent) => {
             // Accumulate locally only — persisted once in finally to avoid
             // one full localStorage write per MCP event during streaming.