diff --git a/core/src/browser/extensions/engines/AIEngine.ts b/core/src/browser/extensions/engines/AIEngine.ts
index a23e8c45e..8853e3334 100644
--- a/core/src/browser/extensions/engines/AIEngine.ts
+++ b/core/src/browser/extensions/engines/AIEngine.ts
@@ -6,6 +6,7 @@ import { EngineManager } from './EngineManager'
 export interface chatCompletionRequestMessage {
   role: 'system' | 'user' | 'assistant' | 'tool'
   content: string | null | Content[] // Content can be a string OR an array of content parts
+  reasoning?: string | null // Some models return reasoning in completed responses
   name?: string
   tool_calls?: any[] // Simplified tool_call_id?: string
 }
diff --git a/web-app/src/hooks/useChat.ts b/web-app/src/hooks/useChat.ts
index 390ebba9e..d77d507c5 100644
--- a/web-app/src/hooks/useChat.ts
+++ b/web-app/src/hooks/useChat.ts
@@ -29,6 +29,10 @@ import { updateSettings } from '@/services/providers'
 import { useContextSizeApproval } from './useModelContextApproval'
 import { useModelLoad } from './useModelLoad'
 import { useGeneralSetting } from './useGeneralSetting'
+import {
+  ReasoningProcessor,
+  extractReasoningFromMessage,
+} from '@/utils/reasoning'
 
 export const useChat = () => {
   const { prompt, setPrompt } = usePrompt()
@@ -285,16 +289,25 @@ export const useChat = () => {
           const toolCalls: ChatCompletionMessageToolCall[] = []
           try {
             if (isCompletionResponse(completion)) {
-              accumulatedText =
-                (completion.choices[0]?.message?.content as string) || ''
-              if (completion.choices[0]?.message?.tool_calls) {
-                toolCalls.push(...completion.choices[0].message.tool_calls)
+              const message = completion.choices[0]?.message
+              accumulatedText = (message?.content as string) || ''
+
+              // Handle reasoning field if there is one
+              const reasoning = extractReasoningFromMessage(message)
+              if (reasoning) {
+                accumulatedText =
+                  `<think>${reasoning}</think>` + accumulatedText
+              }
+
+              if (message?.tool_calls) {
+                toolCalls.push(...message.tool_calls)
               }
             } else {
               // High-throughput scheduler: batch UI updates on rAF (requestAnimationFrame)
               let rafScheduled = false
               let rafHandle: number | undefined
               let pendingDeltaCount = 0
+              const reasoningProcessor = new ReasoningProcessor()
               const scheduleFlush = () => {
                 if (rafScheduled) return
                 rafScheduled = true
@@ -328,7 +341,10 @@ export const useChat = () => {
               }
               const flushIfPending = () => {
                 if (!rafScheduled) return
-                if (typeof cancelAnimationFrame !== 'undefined' && rafHandle !== undefined) {
+                if (
+                  typeof cancelAnimationFrame !== 'undefined' &&
+                  rafHandle !== undefined
+                ) {
                   cancelAnimationFrame(rafHandle)
                 } else if (rafHandle !== undefined) {
                   clearTimeout(rafHandle)
@@ -360,20 +376,30 @@ export const useChat = () => {
                       : (JSON.stringify(part) ?? '')
                   )
                 }
-                const delta = part.choices[0]?.delta?.content || ''
 
                 if (part.choices[0]?.delta?.tool_calls) {
                   extractToolCall(part, currentCall, toolCalls)
                   // Schedule a flush to reflect tool update
                   scheduleFlush()
                 }
-                if (delta) {
-                  accumulatedText += delta
+                const deltaReasoning =
+                  reasoningProcessor.processReasoningChunk(part)
+                if (deltaReasoning) {
+                  accumulatedText += deltaReasoning
+                  pendingDeltaCount += 1
+                  // Schedule flush for reasoning updates
+                  scheduleFlush()
+                }
+                const deltaContent = part.choices[0]?.delta?.content || ''
+                if (deltaContent) {
+                  accumulatedText += deltaContent
                   pendingDeltaCount += 1
                   // Batch UI update on next animation frame
                   scheduleFlush()
                 }
               }
+              // Finalize reasoning (close any open think tags)
+              accumulatedText += reasoningProcessor.finalize()
               // Ensure any pending buffered content is rendered at the end
               flushIfPending()
             }
diff --git a/web-app/src/utils/reasoning.ts b/web-app/src/utils/reasoning.ts
new file mode 100644
index 000000000..f102b8a76
--- /dev/null
+++ b/web-app/src/utils/reasoning.ts
@@ -0,0 +1,74 @@
+import { CompletionResponseChunk } from 'token.js'
+import {
+  chatCompletionChunk,
+  ChatCompletionMessage,
+  chatCompletionRequestMessage,
+} from '@janhq/core'
+
+// Extract reasoning from a message (for completed responses)
+export function extractReasoningFromMessage(
+  message: chatCompletionRequestMessage | ChatCompletionMessage
+): string | null {
+  if (!message) return null
+
+  const extendedMessage = message as chatCompletionRequestMessage
+  return extendedMessage.reasoning || null
+}
+
+// Extract reasoning from a chunk (for streaming responses)
+function extractReasoningFromChunk(
+  chunk: CompletionResponseChunk | chatCompletionChunk
+): string | null {
+  if (!chunk.choices?.[0]?.delta) return null
+
+  const delta = chunk.choices[0].delta as chatCompletionRequestMessage
+  const reasoning = delta.reasoning
+
+  // Return null for falsy values, non-strings, or whitespace-only strings
+  if (!reasoning || typeof reasoning !== 'string' || !reasoning.trim())
+    return null
+
+  return reasoning
+}
+
+// Tracks reasoning state and appends reasoning tokens with proper think tags
+export class ReasoningProcessor {
+  private isReasoningActive = false
+
+  processReasoningChunk(
+    chunk: CompletionResponseChunk | chatCompletionChunk
+  ): string {
+    const reasoning = extractReasoningFromChunk(chunk)
+    const chunkContent = chunk.choices?.[0]?.delta?.content || ''
+
+    // Handle reasoning tokens
+    if (reasoning) {
+      if (!this.isReasoningActive) {
+        this.isReasoningActive = true
+        return '<think>' + reasoning
+      }
+      return reasoning
+    }
+
+    // Handle reasoning end when content starts
+    if (this.isReasoningActive && chunkContent) {
+      this.isReasoningActive = false
+      return '</think>'
+    }
+
+    // No reasoning to process
+    return ''
+  }
+
+  finalize(): string {
+    if (this.isReasoningActive) {
+      this.isReasoningActive = false
+      return '</think>'
+    }
+    return ''
+  }
+
+  isReasoningInProgress(): boolean {
+    return this.isReasoningActive
+  }
+}