diff --git a/CHANGELOG.md b/CHANGELOG.md
index 005bf63c018..4becf532790 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -136,6 +136,7 @@ Docs: https://docs.openclaw.ai
 - Heartbeat: skip wake delivery when the target session lane is already busy so the pending event is retried instead of getting drained too early. (#40526) Thanks @lucky7323.
 - Plugin SDK/context engines: export the missing context-engine result and subagent lifecycle types from `openclaw/plugin-sdk` so context engine plugins can type `ContextEngine` implementations without local workarounds. (#61251) Thanks @DaevMithran.
 - Agents/errors: surface an explicit disk-full message when local session or transcript writes fail with `ENOSPC`/`disk full`, so those runs stop degrading into opaque `NO_REPLY`-style failures. Thanks @vincentkoc.
+- Telegram/reasoning: only create a Telegram reasoning preview lane when the session is explicitly `reasoning:stream`, so hidden `<think>` traces from streamed replies stop surfacing as chat previews on normal sessions. Thanks @vincentkoc.
 
 ## 2026.4.2
 
diff --git a/extensions/telegram/src/bot-message-dispatch.test.ts b/extensions/telegram/src/bot-message-dispatch.test.ts
index 70035d74198..0b7ae188f19 100644
--- a/extensions/telegram/src/bot-message-dispatch.test.ts
+++ b/extensions/telegram/src/bot-message-dispatch.test.ts
@@ -539,6 +539,25 @@ describe("dispatchTelegramMessage draft streaming", () => {
     expect(loadSessionStore).toHaveBeenCalledWith("/tmp/sessions.json", { skipCache: true });
   });
 
+  it("does not expose reasoning preview callbacks unless session reasoning is stream", async () => {
+    let seenReasoningCallback: unknown;
+    const answerDraftStream = createDraftStream(999);
+    createTelegramDraftStream.mockImplementationOnce(() => answerDraftStream);
+    dispatchReplyWithBufferedBlockDispatcher.mockImplementation(async ({ replyOptions }) => {
+      seenReasoningCallback = replyOptions?.onReasoningStream;
+      await replyOptions?.onPartialReply?.({
+        text: "<think>internal chain of thought</think>Visible answer",
+      });
+      return { queuedFinal: false };
+    });
+
+    await dispatchWithContext({ context: createContext(), streamMode: "partial" });
+
+    expect(seenReasoningCallback).toBeUndefined();
+    expect(createTelegramDraftStream).toHaveBeenCalledTimes(1);
+    expect(answerDraftStream.update).toHaveBeenCalledWith("Visible answer");
+  });
+
   it("does not overwrite finalized preview when additional final payloads are sent", async () => {
     const draftStream = createDraftStream(999);
     createTelegramDraftStream.mockReturnValue(draftStream);
diff --git a/extensions/telegram/src/bot-message-dispatch.ts b/extensions/telegram/src/bot-message-dispatch.ts
index 93ab6f0c2ae..ed5d3653fe1 100644
--- a/extensions/telegram/src/bot-message-dispatch.ts
+++ b/extensions/telegram/src/bot-message-dispatch.ts
@@ -213,7 +213,7 @@ export const dispatchTelegramMessage = async ({
   const previewStreamingEnabled = streamMode !== "off";
   const canStreamAnswerDraft =
     previewStreamingEnabled && !accountBlockStreamingEnabled && !forceBlockStreamingForReasoning;
-  const canStreamReasoningDraft = canStreamAnswerDraft || streamReasoningDraft;
+  const canStreamReasoningDraft = streamReasoningDraft;
   const draftReplyToMessageId =
     replyToMode !== "off" && typeof msg.message_id === "number" ? msg.message_id : undefined;
   const draftMinInitialChars = DRAFT_MIN_INITIAL_CHARS;