diff --git a/CHANGELOG.md b/CHANGELOG.md
index 6b726dfef..8fbddabe0 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -24,6 +24,7 @@ Docs: https://docs.openclaw.ai
 - Sessions/Store: canonicalize inbound mixed-case session keys for metadata and route updates, and migrate legacy case-variant entries to a single lowercase key to prevent duplicate sessions and missing TUI/WebUI history. (#9561) Thanks @hillghost86.
 - Telegram/Reactions: soft-fail reaction action errors (policy/token/emoji/API), accept snake_case `message_id`, and fallback to inbound message-id context when explicit `messageId` is omitted so DM reactions stay stable without regeneration loops. (#20236, #21001) Thanks @PeterShanxin and @vincentkoc.
 - Telegram/Polling: scope persisted polling offsets to bot identity and reuse a single awaited runner-stop path on abort/retry, preventing cross-token offset bleed and overlapping pollers during restart/error recovery. (#10850, #11347) Thanks @talhaorak, @anooprdawar, and @vincentkoc.
+- Telegram/Reasoning: when `/reasoning off` is active, suppress reasoning-only delivery segments and block raw fallback resend of suppressed `Reasoning:`/`<think>` text, preventing internal reasoning leakage in legacy sessions while preserving answer delivery. (#24626, #24518)
 - Agents/Reasoning: when model-default thinking is active (for example `thinking=low`), keep auto-reasoning disabled unless explicitly enabled, preventing `Reasoning:` thinking-block leakage in channel replies. (#24335, #24290) thanks @Kay-051.
 - Agents/Reasoning: avoid classifying provider reasoning-required errors as context overflows so these failures no longer trigger compaction-style overflow recovery. (#24593) Thanks @vincentkoc.
 - Agents/Models: codify `agents.defaults.model` / `agents.defaults.imageModel` config-boundary input as `string | {primary,fallbacks}`, split explicit vs effective model resolution, and fix `models status --agent` source attribution so defaults-inherited agents are labeled as `defaults` while runtime selection still honors defaults fallback. (#24210) thanks @bianbiandashen.
diff --git a/src/telegram/bot-message-dispatch.test.ts b/src/telegram/bot-message-dispatch.test.ts
index 5e080e90e..75a8fb6b9 100644
--- a/src/telegram/bot-message-dispatch.test.ts
+++ b/src/telegram/bot-message-dispatch.test.ts
@@ -176,6 +176,15 @@ describe("dispatchTelegramMessage draft streaming", () => {
     });
   }
 
+  function createReasoningStreamContext(): TelegramMessageContext {
+    loadSessionStore.mockReturnValue({
+      s1: { reasoningLevel: "stream" },
+    });
+    return createContext({
+      ctxPayload: { SessionKey: "s1" } as unknown as TelegramMessageContext["ctxPayload"],
+    });
+  }
+
   it("streams drafts in private threads and forwards thread id", async () => {
     const draftStream = createDraftStream();
     createTelegramDraftStream.mockReturnValue(draftStream);
@@ -772,7 +781,7 @@ describe("dispatchTelegramMessage draft streaming", () => {
       deliverReplies.mockResolvedValue({ delivered: true });
       editMessageTelegram.mockResolvedValue({ ok: true, chatId: "123", messageId: "999" });
 
-      await dispatchWithContext({ context: createContext(), streamMode });
+      await dispatchWithContext({ context: createReasoningStreamContext(), streamMode });
 
       expect(reasoningDraftStream.forceNewMessage).toHaveBeenCalledTimes(1);
     },
@@ -809,7 +818,11 @@ describe("dispatchTelegramMessage draft streaming", () => {
     editMessageTelegram.mockResolvedValue({ ok: true, chatId: "123", messageId: "999" });
 
     const bot = createBot();
-    await dispatchWithContext({ context: createContext(), streamMode: "partial", bot });
+    await dispatchWithContext({
+      context: createReasoningStreamContext(),
+      streamMode: "partial",
+      bot,
+    });
 
     expect(reasoningDraftParams?.onSupersededPreview).toBeTypeOf("function");
     const deleteMessageCalls = (
@@ -836,13 +849,13 @@ describe("dispatchTelegramMessage draft streaming", () => {
       );
       deliverReplies.mockResolvedValue({ delivered: true });
 
-      await dispatchWithContext({ context: createContext(), streamMode });
+      await dispatchWithContext({ context: createReasoningStreamContext(), streamMode });
 
       expect(reasoningDraftStream.forceNewMessage).not.toHaveBeenCalled();
     },
   );
 
-  it("does not finalize preview with reasoning payloads before answer payloads", async () => {
+  it("suppresses reasoning-only final payloads when reasoning level is off", async () => {
     setupDraftStreams({ answerMessageId: 999 });
     dispatchReplyWithBufferedBlockDispatcher.mockImplementation(
       async ({ dispatcherOptions, replyOptions }) => {
@@ -860,14 +873,11 @@ describe("dispatchTelegramMessage draft streaming", () => {
 
     await dispatchWithContext({ context: createContext(), streamMode: "partial" });
 
-    // Keep reasoning as its own message.
-    expect(deliverReplies).toHaveBeenCalledTimes(1);
-    expect(deliverReplies).toHaveBeenCalledWith(
+    expect(deliverReplies).not.toHaveBeenCalledWith(
       expect.objectContaining({
         replies: [expect.objectContaining({ text: "Reasoning:\n_step one_" })],
       }),
     );
-    // Finalize preview with the actual answer instead of overwriting with reasoning.
     expect(editMessageTelegram).toHaveBeenCalledTimes(1);
     expect(editMessageTelegram).toHaveBeenCalledWith(
       123,
@@ -877,6 +887,25 @@ describe("dispatchTelegramMessage draft streaming", () => {
     );
   });
 
+  it("does not resend suppressed reasoning-only text through raw fallback", async () => {
+    setupDraftStreams({ answerMessageId: 999 });
+    dispatchReplyWithBufferedBlockDispatcher.mockImplementation(async ({ dispatcherOptions }) => {
+      await dispatcherOptions.deliver({ text: "Reasoning:\n_step one_" }, { kind: "final" });
+      return { queuedFinal: true };
+    });
+    deliverReplies.mockResolvedValue({ delivered: true });
+    editMessageTelegram.mockResolvedValue({ ok: true, chatId: "123", messageId: "999" });
+
+    await dispatchWithContext({ context: createContext(), streamMode: "partial" });
+
+    expect(deliverReplies).not.toHaveBeenCalledWith(
+      expect.objectContaining({
+        replies: [expect.objectContaining({ text: "Reasoning:\n_step one_" })],
+      }),
+    );
+    expect(editMessageTelegram).not.toHaveBeenCalled();
+  });
+
   it("keeps reasoning and answer streaming in separate preview lanes", async () => {
     const { answerDraftStream, reasoningDraftStream } = setupDraftStreams({
       answerMessageId: 999,
@@ -893,7 +922,7 @@ describe("dispatchTelegramMessage draft streaming", () => {
     deliverReplies.mockResolvedValue({ delivered: true });
     editMessageTelegram.mockResolvedValue({ ok: true, chatId: "123", messageId: "999" });
 
-    await dispatchWithContext({ context: createContext(), streamMode: "partial" });
+    await dispatchWithContext({ context: createReasoningStreamContext(), streamMode: "partial" });
 
     expect(reasoningDraftStream.update).toHaveBeenCalledWith("Reasoning:\n_Working on it..._");
     expect(answerDraftStream.update).toHaveBeenCalledWith("Checking the directory...");
@@ -913,7 +942,7 @@ describe("dispatchTelegramMessage draft streaming", () => {
     deliverReplies.mockResolvedValue({ delivered: true });
     editMessageTelegram.mockResolvedValue({ ok: true, chatId: "123", messageId: "999" });
 
-    await dispatchWithContext({ context: createContext(), streamMode: "partial" });
+    await dispatchWithContext({ context: createReasoningStreamContext(), streamMode: "partial" });
 
     expect(editMessageTelegram).not.toHaveBeenCalled();
     expect(deliverReplies).toHaveBeenCalledWith(
@@ -955,7 +984,7 @@ describe("dispatchTelegramMessage draft streaming", () => {
       deliverReplies.mockResolvedValue({ delivered: true });
       editMessageTelegram.mockResolvedValue({ ok: true, chatId: "123", messageId: "111" });
 
-      await dispatchWithContext({ context: createContext(), streamMode });
+      await dispatchWithContext({ context: createReasoningStreamContext(), streamMode });
 
       expect(reasoningDraftStream.forceNewMessage).not.toHaveBeenCalled();
       expect(editMessageTelegram).toHaveBeenCalledWith(
@@ -990,7 +1019,7 @@ describe("dispatchTelegramMessage draft streaming", () => {
     deliverReplies.mockResolvedValue({ delivered: true });
     editMessageTelegram.mockResolvedValue({ ok: true, chatId: "123", messageId: "999" });
 
-    await dispatchWithContext({ context: createContext(), streamMode: "partial" });
+    await dispatchWithContext({ context: createReasoningStreamContext(), streamMode: "partial" });
 
     expect(editMessageTelegram).toHaveBeenNthCalledWith(1, 123, 999, "3", expect.any(Object));
     expect(editMessageTelegram).toHaveBeenNthCalledWith(
@@ -1028,7 +1057,7 @@ describe("dispatchTelegramMessage draft streaming", () => {
     deliverReplies.mockResolvedValue({ delivered: true });
     editMessageTelegram.mockResolvedValue({ ok: true, chatId: "123", messageId: "999" });
 
-    await dispatchWithContext({ context: createContext(), streamMode: "partial" });
+    await dispatchWithContext({ context: createReasoningStreamContext(), streamMode: "partial" });
 
     expect(reasoningDraftStream.update).toHaveBeenCalledWith(
       "Reasoning:\n_Counting letters in strawberry_",
@@ -1060,7 +1089,7 @@ describe("dispatchTelegramMessage draft streaming", () => {
     deliverReplies.mockResolvedValue({ delivered: true });
     editMessageTelegram.mockResolvedValue({ ok: true, chatId: "123", messageId: "999" });
 
-    await dispatchWithContext({ context: createContext(), streamMode: "partial" });
+    await dispatchWithContext({ context: createReasoningStreamContext(), streamMode: "partial" });
 
     expect(reasoningDraftStream.update).toHaveBeenCalledWith(
       "Reasoning:\n_Counting letters in strawberry_",
@@ -1096,7 +1125,7 @@ describe("dispatchTelegramMessage draft streaming", () => {
     deliverReplies.mockResolvedValue({ delivered: true });
     editMessageTelegram.mockResolvedValue({ ok: true, chatId: "123", messageId: "999" });
 
-    await dispatchWithContext({ context: createContext(), streamMode: "partial" });
+    await dispatchWithContext({ context: createReasoningStreamContext(), streamMode: "partial" });
 
     expect(reasoningDraftStream.update).toHaveBeenCalledWith(
       "Reasoning:\n_Word: strawberry. r appears at 3, 8, 9._",
@@ -1127,7 +1156,7 @@ describe("dispatchTelegramMessage draft streaming", () => {
     deliverReplies.mockResolvedValue({ delivered: true });
     editMessageTelegram.mockResolvedValue({ ok: true, chatId: "123", messageId: "999" });
 
-    await dispatchWithContext({ context: createContext(), streamMode: "partial" });
+    await dispatchWithContext({ context: createReasoningStreamContext(), streamMode: "partial" });
 
     expect(editMessageTelegram).toHaveBeenNthCalledWith(
       1,
diff --git a/src/telegram/bot-message-dispatch.ts b/src/telegram/bot-message-dispatch.ts
index 443555fdd..7dd0c4845 100644
--- a/src/telegram/bot-message-dispatch.ts
+++ b/src/telegram/bot-message-dispatch.ts
@@ -202,16 +202,25 @@ export const dispatchTelegramMessage = async ({
   let splitReasoningOnNextStream = false;
   const reasoningStepState = createTelegramReasoningStepState();
   type SplitLaneSegment = { lane: LaneName; text: string };
-  const splitTextIntoLaneSegments = (text?: string): SplitLaneSegment[] => {
+  type SplitLaneSegmentsResult = {
+    segments: SplitLaneSegment[];
+    suppressedReasoningOnly: boolean;
+  };
+  const splitTextIntoLaneSegments = (text?: string): SplitLaneSegmentsResult => {
     const split = splitTelegramReasoningText(text);
     const segments: SplitLaneSegment[] = [];
-    if (split.reasoningText) {
+    const suppressReasoning = resolvedReasoningLevel === "off";
+    if (split.reasoningText && !suppressReasoning) {
       segments.push({ lane: "reasoning", text: split.reasoningText });
     }
     if (split.answerText) {
       segments.push({ lane: "answer", text: split.answerText });
     }
-    return segments;
+    return {
+      segments,
+      suppressedReasoningOnly:
+        Boolean(split.reasoningText) && suppressReasoning && !split.answerText,
+    };
   };
   const resetDraftLaneState = (lane: DraftLaneState) => {
     lane.lastPartialText = "";
@@ -241,7 +250,8 @@ export const dispatchTelegramMessage = async ({
     laneStream.update(text);
   };
   const ingestDraftLaneSegments = (text: string | undefined) => {
-    for (const segment of splitTextIntoLaneSegments(text)) {
+    const split = splitTextIntoLaneSegments(text);
+    for (const segment of split.segments) {
       if (segment.lane === "reasoning") {
         reasoningStepState.noteReasoningHint();
         reasoningStepState.noteReasoningDelivered();
@@ -418,7 +428,8 @@ export const dispatchTelegramMessage = async ({
           const previewButtons = (
             payload.channelData?.telegram as { buttons?: TelegramInlineButtons } | undefined
           )?.buttons;
-          const segments = splitTextIntoLaneSegments(payload.text);
+          const split = splitTextIntoLaneSegments(payload.text);
+          const segments = split.segments;
           const hasMedia = Boolean(payload.mediaUrl) || (payload.mediaUrls?.length ?? 0) > 0;
 
           const flushBufferedFinalAnswer = async () => {
@@ -478,6 +489,17 @@ export const dispatchTelegramMessage = async ({
           if (segments.length > 0) {
             return;
           }
+          if (split.suppressedReasoningOnly) {
+            if (hasMedia) {
+              const payloadWithoutSuppressedReasoning =
+                typeof payload.text === "string" ? { ...payload, text: "" } : payload;
+              await sendPayload(payloadWithoutSuppressedReasoning);
+            }
+            if (info.kind === "final") {
+              await flushBufferedFinalAnswer();
+            }
+            return;
+          }
 
           if (info.kind === "final") {
             await answerLane.stream?.stop();