diff --git a/src/main/apiServer/middleware/gateway.ts b/src/main/apiServer/middleware/gateway.ts
index 2de7a6a726..a4677c5523 100644
--- a/src/main/apiServer/middleware/gateway.ts
+++ b/src/main/apiServer/middleware/gateway.ts
@@ -5,219 +5,25 @@
  * - Model group routing: /{groupName}/v1/... routes use the group's configured model
  * - Endpoint access control based on enabledEndpoints configuration
  * - Model injection for simplified external app integration
+ *
+ * For assistant mode, this middleware only sets model = "assistant:{assistantId}"
+ * The actual assistant config resolution and parameter overrides happen in ProxyStreamService.
  */
 
-import { buildFunctionCallToolName } from '@main/utils/mcp'
-import type { MCPTool } from '@types'
 import type { NextFunction, Request, Response } from 'express'
 
 import { loggerService } from '../../services/LoggerService'
-import { reduxService } from '../../services/ReduxService'
 import { config } from '../config'
-import { mcpApiService } from '../services/mcp'
-import { getMCPServersFromRedux } from '../utils/mcp'
 
 const logger = loggerService.withContext('GatewayMiddleware')
 
-type AssistantConfig = {
-  id: string
-  name: string
-  prompt?: string
-  model?: { id: string; provider: string }
-  defaultModel?: { id: string; provider: string }
-  settings?: {
-    streamOutput?: boolean
-    enableTemperature?: boolean
-    temperature?: number
-    enableTopP?: boolean
-    topP?: number
-    enableMaxTokens?: boolean
-    maxTokens?: number
-  }
-  mcpServers?: Array<{ id: string }>
-  allowed_tools?: string[]
-}
-
-type ToolDefinition = {
-  name: string
-  description?: string
-  inputSchema: MCPTool['inputSchema']
-}
-
-const getEndpointFormat = (endpoint: string): 'openai' | 'anthropic' | 'responses' | null => {
-  if (endpoint.startsWith('/v1/chat/completions')) return 'openai'
-  if (endpoint.startsWith('/v1/messages')) return 'anthropic'
-  if (endpoint.startsWith('/v1/responses')) return 'responses'
-  return null
-}
-
-const buildAssistantModelId = (assistant: AssistantConfig): string | null => {
-  const model = assistant.model ?? assistant.defaultModel
-  if (!model?.provider || !model?.id) {
-    return null
-  }
-  return `${model.provider}:${model.id}`
-}
-
-const applyAssistantMessageOverrides = (
-  body: Record<string, any>,
-  assistant: AssistantConfig,
-  format: 'openai' | 'anthropic' | 'responses'
-): Record<string, any> => {
-  const nextBody = { ...body }
-  const prompt = assistant.prompt ?? ''
-
-  if (format === 'openai') {
-    const messages = Array.isArray(nextBody.messages) ? nextBody.messages : []
-    const filtered = messages.filter((message) => message?.role !== 'system' && message?.role !== 'developer')
-    if (prompt.trim().length > 0) {
-      filtered.unshift({ role: 'system', content: prompt })
-    }
-    nextBody.messages = filtered
-  } else if (format === 'responses') {
-    nextBody.instructions = prompt
-  } else {
-    nextBody.system = prompt
-  }
-
-  return nextBody
-}
-
-const applyAssistantParameterOverrides = (
-  body: Record<string, any>,
-  assistant: AssistantConfig,
-  format: 'openai' | 'anthropic' | 'responses'
-): Record<string, any> => {
-  const nextBody = { ...body }
-  const settings = assistant.settings ?? {}
-
-  if (typeof settings.streamOutput === 'boolean') {
-    nextBody.stream = settings.streamOutput
-  }
-
-  if (settings.enableTemperature && typeof settings.temperature === 'number') {
-    nextBody.temperature = settings.temperature
-  } else if ('temperature' in nextBody) {
-    delete nextBody.temperature
-  }
-
-  if (settings.enableTopP && typeof settings.topP === 'number') {
-    nextBody.top_p = settings.topP
-  } else if ('top_p' in nextBody) {
-    delete nextBody.top_p
-  }
-
-  if (settings.enableMaxTokens && typeof settings.maxTokens === 'number') {
-    if (format === 'responses') {
-      nextBody.max_output_tokens = settings.maxTokens
-      delete nextBody.max_tokens
-    } else {
-      nextBody.max_tokens = settings.maxTokens
-      if ('max_output_tokens' in nextBody) {
-        delete nextBody.max_output_tokens
-      }
-    }
-  } else {
-    if ('max_tokens' in nextBody) {
-      delete nextBody.max_tokens
-    }
-    if ('max_output_tokens' in nextBody) {
-      delete nextBody.max_output_tokens
-    }
-  }
-
-  delete nextBody.tool_choice
-
-  return nextBody
-}
-
-const mapToolsForOpenAI = (tools: ToolDefinition[]) =>
-  tools.map((toolDef) => ({
-    type: 'function',
-    function: {
-      name: toolDef.name,
-      description: toolDef.description || '',
-      parameters: toolDef.inputSchema
-    }
-  }))
-
-const mapToolsForResponses = (tools: ToolDefinition[]) =>
-  tools.map((toolDef) => ({
-    type: 'function',
-    name: toolDef.name,
-    description: toolDef.description || '',
-    parameters: toolDef.inputSchema
-  }))
-
-const mapToolsForAnthropic = (tools: ToolDefinition[]) =>
-  tools.map((toolDef) => ({
-    name: toolDef.name,
-    description: toolDef.description || '',
-    input_schema: toolDef.inputSchema
-  }))
-
-const buildAssistantTools = async (assistant: AssistantConfig): Promise<ToolDefinition[]> => {
-  const serverIds = assistant.mcpServers?.map((server) => server.id).filter(Boolean) ?? []
-  if (serverIds.length === 0) {
-    return []
-  }
-
-  const allowedTools = Array.isArray(assistant.allowed_tools) ? new Set(assistant.allowed_tools) : null
-  const servers = await getMCPServersFromRedux()
-  const tools: ToolDefinition[] = []
-
-  for (const serverId of serverIds) {
-    const server = servers.find((entry) => entry.id === serverId)
-    if (!server || !server.isActive) {
-      continue
-    }
-
-    const info = await mcpApiService.getServerInfo(serverId)
-    if (!info?.tools || !Array.isArray(info.tools)) {
-      continue
-    }
-
-    for (const tool of info.tools as Array<{
-      name: string
-      description?: string
-      inputSchema?: MCPTool['inputSchema']
-    }>) {
-      if (!tool?.name || !tool.inputSchema) {
-        continue
-      }
-
-      if (server.disabledTools?.includes(tool.name)) {
-        continue
-      }
-
-      const toolName = buildFunctionCallToolName(info.name, tool.name)
-      if (allowedTools && !allowedTools.has(toolName)) {
-        continue
-      }
-
-      tools.push({
-        name: toolName,
-        description: tool.description,
-        inputSchema: tool.inputSchema
-      })
-    }
-  }
-
-  return tools
-}
-
-const resolveAssistantById = async (assistantId: string): Promise<AssistantConfig | null> => {
-  const assistants = (await reduxService.select('state.assistants.assistants')) as AssistantConfig[] | null
-  return assistants?.find((assistant) => assistant.id === assistantId) ?? null
-}
-
 /**
  * Gateway middleware for model group routing
  *
  * This middleware:
  * 1. Extracts group name from URL path if present
  * 2. Looks up the group by matching name directly
- * 3. Injects the group's model into the request
+ * 3. Injects the group's model into the request (or assistant ID for assistant mode)
  * 4. Checks if the endpoint is enabled
  */
 export const gatewayMiddleware = async (req: Request, res: Response, next: NextFunction): Promise<void> => {
@@ -240,69 +46,18 @@ export const gatewayMiddleware = async (req: Request, res: Response, next: NextF
         return
       }
 
-      const endpoint = req.path.startsWith('/') ? req.path : `/${req.path}`
-      const endpointFormat = getEndpointFormat(endpoint)
-
       if (group.mode === 'assistant' && group.assistantId) {
-        if (!endpointFormat) {
-          res.status(400).json({
-            error: {
-              type: 'invalid_request_error',
-              message: `Unsupported endpoint ${endpoint}`
-            }
-          })
-          return
-        }
-
-        const assistant = await resolveAssistantById(group.assistantId)
-        if (!assistant) {
-          res.status(400).json({
-            error: {
-              type: 'invalid_request_error',
-              message: `Assistant '${group.assistantId}' not found`
-            }
-          })
-          return
-        }
-
-        const modelId = buildAssistantModelId(assistant)
-        if (!modelId) {
-          res.status(400).json({
-            error: {
-              type: 'invalid_request_error',
-              message: `Assistant '${group.assistantId}' is missing a model`
-            }
-          })
-          return
-        }
-
-        let nextBody = {
+        req.body = {
           ...req.body,
-          model: modelId
+          model: `assistant:${group.assistantId}`
         }
 
-        nextBody = applyAssistantMessageOverrides(nextBody, assistant, endpointFormat)
-        nextBody = applyAssistantParameterOverrides(nextBody, assistant, endpointFormat)
-
-        const tools = await buildAssistantTools(assistant)
-        if (endpointFormat === 'openai') {
-          nextBody.tools = tools.length > 0 ? mapToolsForOpenAI(tools) : undefined
-        } else if (endpointFormat === 'responses') {
-          nextBody.tools = tools.length > 0 ? mapToolsForResponses(tools) : undefined
-        } else {
-          nextBody.tools = tools.length > 0 ? mapToolsForAnthropic(tools) : undefined
-        }
-
-        req.body = nextBody
-
-        logger.debug('Injected assistant preset into request', {
+        logger.debug('Using assistant mode', {
           groupName,
-          assistantId: assistant.id,
-          model: modelId,
-          toolCount: tools.length
+          assistantId: group.assistantId
         })
       } else {
-        // Inject the group's model into the request
+        // Model mode: inject the group's model into the request
         req.body = {
           ...req.body,
           model: `${group.providerId}:${group.modelId}`
diff --git a/src/main/apiServer/services/ProxyStreamService.ts b/src/main/apiServer/services/ProxyStreamService.ts
index d5387c9f23..9d4d67bd8e 100644
--- a/src/main/apiServer/services/ProxyStreamService.ts
+++ b/src/main/apiServer/services/ProxyStreamService.ts
@@ -12,6 +12,7 @@ import { loggerService } from '@logger'
 import { generateSignature as cherryaiGenerateSignature } from '@main/integration/cherryai'
 import anthropicService from '@main/services/AnthropicService'
 import copilotService from '@main/services/CopilotService'
+import mcpService from '@main/services/MCPService'
 import { reduxService } from '@main/services/ReduxService'
 import {
   type AiSdkConfig,
@@ -25,6 +26,7 @@ import {
 import { COPILOT_DEFAULT_HEADERS } from '@shared/aiCore/constant'
 import type { MinimalProvider } from '@shared/types'
 import { defaultAppHeaders } from '@shared/utils'
+import type { Assistant, MCPTool } from '@types'
 import type { Provider } from '@types'
 import type { Provider as AiSdkProvider } from 'ai'
 import { simulateStreamingMiddleware, stepCountIs, wrapLanguageModel } from 'ai'
@@ -44,6 +46,108 @@ initializeSharedProviders({
   error: (message, error) => logger.error(message, error)
 })
 
+/**
+ * Resolve assistant configuration by ID from Redux store
+ */
+async function resolveAssistantById(assistantId: string): Promise<Assistant | null> {
+  const assistants = (await reduxService.select('state.assistants.assistants')) as Assistant[] | null
+  return assistants?.find((assistant) => assistant.id === assistantId) ?? null
+}
+
+/**
+ * Get provider by ID from Redux store
+ */
+async function getProviderById(providerId: string): Promise<Provider | null> {
+  const providers = (await reduxService.select('state.llm.providers')) as Provider[] | null
+  return providers?.find((p) => p.id === providerId) ?? null
+}
+
+/**
+ * Build MCP tools for assistant using Promise.allSettled for concurrent fetching
+ */
+async function buildAssistantTools(assistant: Assistant): Promise<MCPTool[]> {
+  const servers = assistant.mcpServers ?? []
+  if (servers.length === 0) {
+    return []
+  }
+
+  logger.debug('Fetching MCP tools for assistant', { serverCount: servers.length })
+
+  const results = await Promise.allSettled(servers.map((server) => mcpService.listTools(null as never, server)))
+
+  const tools: MCPTool[] = []
+  for (let i = 0; i < results.length; i++) {
+    const result = results[i]
+    if (result.status === 'fulfilled') {
+      tools.push(...result.value)
+    } else {
+      logger.warn('Failed to fetch tools from MCP server', {
+        serverId: servers[i].id,
+        serverName: servers[i].name,
+        error: result.reason
+      })
+    }
+  }
+
+  logger.debug('Built MCP tools for assistant', { toolCount: tools.length })
+  return tools
+}
+
+/**
+ * Apply assistant overrides to request params based on input format
+ *
+ * Only handles format-specific field differences:
+ * - System prompt: system (Anthropic) / instructions (Responses) / messages[0] (OpenAI)
+ * - MaxTokens: max_tokens / max_output_tokens
+ *
+ * AI SDK's MessageConverter handles the rest of the parameter transformations.
+ */
+function applyAssistantOverrides(params: InputParams, assistant: Assistant, inputFormat: InputFormat): InputParams {
+  const nextParams = { ...params } as Record<string, unknown>
+  const settings = assistant.settings ?? {}
+  const prompt = assistant.prompt?.trim() ?? ''
+
+  // 1. System prompt (format-specific field)
+  if (prompt.length > 0) {
+    if (inputFormat === 'openai') {
+      const messages = Array.isArray(nextParams.messages) ? [...nextParams.messages] : []
+      const filtered = messages.filter((m: { role?: string }) => m?.role !== 'system' && m?.role !== 'developer')
+      filtered.unshift({ role: 'system', content: prompt })
+      nextParams.messages = filtered
+    } else if (inputFormat === 'openai-responses') {
+      nextParams.instructions = prompt
+    } else {
+      nextParams.system = prompt
+    }
+  }
+
+  // 2. Stream output
+  if (typeof settings.streamOutput === 'boolean') {
+    nextParams.stream = settings.streamOutput
+  }
+
+  // 3. Temperature
+  if (settings.enableTemperature && typeof settings.temperature === 'number') {
+    nextParams.temperature = settings.temperature
+  }
+
+  // 4. TopP
+  if (settings.enableTopP && typeof settings.topP === 'number') {
+    nextParams.top_p = settings.topP
+  }
+
+  // 5. MaxTokens (format-specific field name)
+  if (settings.enableMaxTokens && typeof settings.maxTokens === 'number') {
+    if (inputFormat === 'openai-responses') {
+      nextParams.max_output_tokens = settings.maxTokens
+    } else {
+      nextParams.max_tokens = settings.maxTokens
+    }
+  }
+
+  return nextParams as InputParams
+}
+
 /**
  * Middleware type alias
  */
@@ -59,8 +163,8 @@ type InputParams = InputParamsMap[InputFormat]
  */
 export interface MessageConfig {
   response: Response
-  provider: Provider
-  modelId: string
+  provider?: Provider
+  modelId?: string
   params: InputParams
   inputFormat?: InputFormat
   outputFormat?: OutputFormat
@@ -297,9 +401,6 @@ async function executeStream(config: ExecuteStreamConfig): Promise<{
 export async function processMessage(config: MessageConfig): Promise<void> {
   const {
     response,
-    provider,
-    modelId,
-    params,
     inputFormat = 'anthropic',
     outputFormat = 'anthropic',
     onError,
@@ -308,6 +409,83 @@ export async function processMessage(config: MessageConfig): Promise<void> {
     plugins = []
   } = config
 
+  let { provider, modelId, params } = config
+
+  // Check if model starts with "assistant:" prefix - auto-detect assistant mode
+  const modelFromParams = 'model' in params ? (params as { model?: string }).model : undefined
+  const isAssistantMode = modelFromParams?.startsWith('assistant:')
+
+  // Handle assistant mode: resolve assistant config and apply overrides
+  if (isAssistantMode) {
+    const assistantId = modelFromParams!.slice('assistant:'.length)
+
+    const assistant = await resolveAssistantById(assistantId)
+    if (!assistant) {
+      throw new Error(`Assistant '${assistantId}' not found`)
+    }
+
+    // Get model from assistant config
+    const assistantModel = assistant.model ?? assistant.defaultModel
+    if (!assistantModel?.provider || !assistantModel?.id) {
+      throw new Error(`Assistant '${assistantId}' is missing a model configuration`)
+    }
+
+    // Resolve provider
+    const resolvedProvider = await getProviderById(assistantModel.provider)
+    if (!resolvedProvider) {
+      throw new Error(`Provider '${assistantModel.provider}' not found for assistant '${assistantId}'`)
+    }
+
+    provider = resolvedProvider
+    modelId = assistantModel.id
+
+    // Apply assistant overrides to params
+    params = applyAssistantOverrides(params, assistant, inputFormat)
+
+    // Build and inject MCP tools
+    const assistantTools = await buildAssistantTools(assistant)
+    if (assistantTools.length > 0) {
+      // Inject tools based on input format
+      const paramsWithTools = { ...params } as Record<string, unknown>
+      if (inputFormat === 'openai') {
+        paramsWithTools.tools = assistantTools.map((tool) => ({
+          type: 'function',
+          function: {
+            name: tool.name,
+            description: tool.description || '',
+            parameters: tool.inputSchema
+          }
+        }))
+      } else if (inputFormat === 'openai-responses') {
+        paramsWithTools.tools = assistantTools.map((tool) => ({
+          type: 'function',
+          name: tool.name,
+          description: tool.description || '',
+          parameters: tool.inputSchema
+        }))
+      } else {
+        // anthropic
+        paramsWithTools.tools = assistantTools.map((tool) => ({
+          name: tool.name,
+          description: tool.description || '',
+          input_schema: tool.inputSchema
+        }))
+      }
+      params = paramsWithTools as InputParams
+    }
+
+    logger.debug('Resolved assistant config', {
+      assistantId,
+      providerId: provider.id,
+      modelId,
+      toolCount: assistantTools.length
+    })
+  }
+
+  if (!provider || !modelId) {
+    throw new Error('Provider and modelId are required when not using assistant mode')
+  }
+
   const isStreaming = 'stream' in params && params.stream === true
 
   logger.info(`Starting ${isStreaming ? 'streaming' : 'non-streaming'} message`, {
@@ -317,7 +495,8 @@ export async function processMessage(config: MessageConfig): Promise<void> {
     inputFormat,
     outputFormat,
     middlewareCount: middlewares.length,
-    pluginCount: plugins.length
+    pluginCount: plugins.length,
+    assistantMode: isAssistantMode
   })
 
   // Create abort controller with timeout
diff --git a/src/renderer/src/i18n/locales/zh-tw.json b/src/renderer/src/i18n/locales/zh-tw.json
index bd13791879..b929b47dee 100644
--- a/src/renderer/src/i18n/locales/zh-tw.json
+++ b/src/renderer/src/i18n/locales/zh-tw.json
@@ -353,6 +353,13 @@
         "description": "[to be translated]:Create model groups with unique URLs for different provider/model combinations",
         "empty": "[to be translated]:No model groups configured. Click 'Add Group' to create one.",
         "label": "[to be translated]:Model Groups",
+        "mode": {
+          "assistant": "[to be translated]:Assistant Preset",
+          "assistantHint": "[to be translated]:Assistant preset overrides request parameters.",
+          "assistantPlaceholder": "[to be translated]:Select assistant",
+          "label": "[to be translated]:Mode",
+          "model": "[to be translated]:Direct Model"
+        },
         "namePlaceholder": "[to be translated]:Group name"
       },
       "networkAccess": {
diff --git a/src/renderer/src/i18n/translate/de-de.json b/src/renderer/src/i18n/translate/de-de.json
index ca1649ac28..b818b2fd7f 100644
--- a/src/renderer/src/i18n/translate/de-de.json
+++ b/src/renderer/src/i18n/translate/de-de.json
@@ -353,6 +353,13 @@
         "description": "[to be translated]:Create model groups with unique URLs for different provider/model combinations",
         "empty": "[to be translated]:No model groups configured. Click 'Add Group' to create one.",
         "label": "[to be translated]:Model Groups",
+        "mode": {
+          "assistant": "[to be translated]:Assistant Preset",
+          "assistantHint": "[to be translated]:Assistant preset overrides request parameters.",
+          "assistantPlaceholder": "[to be translated]:Select assistant",
+          "label": "[to be translated]:Mode",
+          "model": "[to be translated]:Direct Model"
+        },
         "namePlaceholder": "[to be translated]:Group name"
       },
       "networkAccess": {
@@ -1350,6 +1357,7 @@
     "backup": {
       "file_format": "Backup-Dateiformat fehlerhaft"
     },
+    "base64DataTruncated": "[to be translated]:Base64 image data truncated, size",
     "boundary": {
       "default": {
         "devtools": "Debug-Panel öffnen",
@@ -1430,6 +1438,8 @@
     "text": "Text",
     "toolInput": "Tool-Eingabe",
     "toolName": "Tool-Name",
+    "truncated": "[to be translated]:Data truncated, original size",
+    "truncatedBadge": "[to be translated]:Truncated",
     "unknown": "Unbekannter Fehler",
     "usage": "Nutzung",
     "user_message_not_found": "Ursprüngliche Benutzernachricht nicht gefunden",
diff --git a/src/renderer/src/i18n/translate/el-gr.json b/src/renderer/src/i18n/translate/el-gr.json
index 366d9e6314..55fe470505 100644
--- a/src/renderer/src/i18n/translate/el-gr.json
+++ b/src/renderer/src/i18n/translate/el-gr.json
@@ -353,6 +353,13 @@
         "description": "[to be translated]:Create model groups with unique URLs for different provider/model combinations",
         "empty": "[to be translated]:No model groups configured. Click 'Add Group' to create one.",
         "label": "[to be translated]:Model Groups",
+        "mode": {
+          "assistant": "[to be translated]:Assistant Preset",
+          "assistantHint": "[to be translated]:Assistant preset overrides request parameters.",
+          "assistantPlaceholder": "[to be translated]:Select assistant",
+          "label": "[to be translated]:Mode",
+          "model": "[to be translated]:Direct Model"
+        },
         "namePlaceholder": "[to be translated]:Group name"
       },
       "networkAccess": {
@@ -1350,6 +1357,7 @@
     "backup": {
       "file_format": "Λάθος μορφή αρχείου που επιστρέφεται"
     },
+    "base64DataTruncated": "[to be translated]:Base64 image data truncated, size",
     "boundary": {
       "default": {
         "devtools": "Άνοιγμα πίνακα αποσφαλμάτωσης",
@@ -1430,6 +1438,8 @@
     "text": "κείμενο",
     "toolInput": "εισαγωγή εργαλείου",
     "toolName": "Όνομα εργαλείου",
+    "truncated": "[to be translated]:Data truncated, original size",
+    "truncatedBadge": "[to be translated]:Truncated",
     "unknown": "Άγνωστο σφάλμα",
     "usage": "δοσολογία",
     "user_message_not_found": "Αδυναμία εύρεσης της αρχικής μηνύματος χρήστη",
diff --git a/src/renderer/src/i18n/translate/es-es.json b/src/renderer/src/i18n/translate/es-es.json
index 843d377aae..a9bee3befa 100644
--- a/src/renderer/src/i18n/translate/es-es.json
+++ b/src/renderer/src/i18n/translate/es-es.json
@@ -353,6 +353,13 @@
         "description": "[to be translated]:Create model groups with unique URLs for different provider/model combinations",
         "empty": "[to be translated]:No model groups configured. Click 'Add Group' to create one.",
         "label": "[to be translated]:Model Groups",
+        "mode": {
+          "assistant": "[to be translated]:Assistant Preset",
+          "assistantHint": "[to be translated]:Assistant preset overrides request parameters.",
+          "assistantPlaceholder": "[to be translated]:Select assistant",
+          "label": "[to be translated]:Mode",
+          "model": "[to be translated]:Direct Model"
+        },
         "namePlaceholder": "[to be translated]:Group name"
       },
       "networkAccess": {
@@ -1350,6 +1357,7 @@
     "backup": {
       "file_format": "Formato de archivo de copia de seguridad incorrecto"
     },
+    "base64DataTruncated": "[to be translated]:Base64 image data truncated, size",
     "boundary": {
       "default": {
         "devtools": "Abrir el panel de depuración",
@@ -1430,6 +1438,8 @@
     "text": "Texto",
     "toolInput": "Herramienta de entrada",
     "toolName": "Nombre de la herramienta",
+    "truncated": "[to be translated]:Data truncated, original size",
+    "truncatedBadge": "[to be translated]:Truncated",
     "unknown": "Error desconocido",
     "usage": "Cantidad de uso",
     "user_message_not_found": "No se pudo encontrar el mensaje original del usuario",
diff --git a/src/renderer/src/i18n/translate/fr-fr.json b/src/renderer/src/i18n/translate/fr-fr.json
index 3e180caac1..8c3fc5d659 100644
--- a/src/renderer/src/i18n/translate/fr-fr.json
+++ b/src/renderer/src/i18n/translate/fr-fr.json
@@ -353,6 +353,13 @@
         "description": "[to be translated]:Create model groups with unique URLs for different provider/model combinations",
         "empty": "[to be translated]:No model groups configured. Click 'Add Group' to create one.",
         "label": "[to be translated]:Model Groups",
+        "mode": {
+          "assistant": "[to be translated]:Assistant Preset",
+          "assistantHint": "[to be translated]:Assistant preset overrides request parameters.",
+          "assistantPlaceholder": "[to be translated]:Select assistant",
+          "label": "[to be translated]:Mode",
+          "model": "[to be translated]:Direct Model"
+        },
         "namePlaceholder": "[to be translated]:Group name"
       },
       "networkAccess": {
@@ -1350,6 +1357,7 @@
     "backup": {
       "file_format": "Le format du fichier de sauvegarde est incorrect"
     },
+    "base64DataTruncated": "[to be translated]:Base64 image data truncated, size",
     "boundary": {
       "default": {
         "devtools": "Ouvrir le panneau de débogage",
@@ -1430,6 +1438,8 @@
     "text": "texte",
     "toolInput": "entrée de l'outil",
     "toolName": "Nom de l'outil",
+    "truncated": "[to be translated]:Data truncated, original size",
+    "truncatedBadge": "[to be translated]:Truncated",
     "unknown": "Неизвестная ошибка",
     "usage": "Quantité",
     "user_message_not_found": "Impossible de trouver le message d'utilisateur original",
diff --git a/src/renderer/src/i18n/translate/ja-jp.json b/src/renderer/src/i18n/translate/ja-jp.json
index a41d9a6789..50a92632aa 100644
--- a/src/renderer/src/i18n/translate/ja-jp.json
+++ b/src/renderer/src/i18n/translate/ja-jp.json
@@ -353,6 +353,13 @@
         "description": "[to be translated]:Create model groups with unique URLs for different provider/model combinations",
         "empty": "[to be translated]:No model groups configured. Click 'Add Group' to create one.",
         "label": "[to be translated]:Model Groups",
+        "mode": {
+          "assistant": "[to be translated]:Assistant Preset",
+          "assistantHint": "[to be translated]:Assistant preset overrides request parameters.",
+          "assistantPlaceholder": "[to be translated]:Select assistant",
+          "label": "[to be translated]:Mode",
+          "model": "[to be translated]:Direct Model"
+        },
         "namePlaceholder": "[to be translated]:Group name"
       },
       "networkAccess": {
@@ -1350,6 +1357,7 @@
     "backup": {
       "file_format": "バックアップファイルの形式エラー"
     },
+    "base64DataTruncated": "[to be translated]:Base64 image data truncated, size",
     "boundary": {
       "default": {
         "devtools": "デバッグパネルを開く",
@@ -1430,6 +1438,8 @@
     "text": "テキスト",
     "toolInput": "ツール入力",
     "toolName": "ツール名",
+    "truncated": "[to be translated]:Data truncated, original size",
+    "truncatedBadge": "[to be translated]:Truncated",
     "unknown": "不明なエラー",
     "usage": "用量",
     "user_message_not_found": "元のユーザーメッセージを見つけることができませんでした",
diff --git a/src/renderer/src/i18n/translate/pt-pt.json b/src/renderer/src/i18n/translate/pt-pt.json
index 8e227f630b..6423a270f9 100644
--- a/src/renderer/src/i18n/translate/pt-pt.json
+++ b/src/renderer/src/i18n/translate/pt-pt.json
@@ -353,6 +353,13 @@
         "description": "[to be translated]:Create model groups with unique URLs for different provider/model combinations",
         "empty": "[to be translated]:No model groups configured. Click 'Add Group' to create one.",
         "label": "[to be translated]:Model Groups",
+        "mode": {
+          "assistant": "[to be translated]:Assistant Preset",
+          "assistantHint": "[to be translated]:Assistant preset overrides request parameters.",
+          "assistantPlaceholder": "[to be translated]:Select assistant",
+          "label": "[to be translated]:Mode",
+          "model": "[to be translated]:Direct Model"
+        },
         "namePlaceholder": "[to be translated]:Group name"
       },
       "networkAccess": {
@@ -1350,6 +1357,7 @@
     "backup": {
       "file_format": "Formato do arquivo de backup está incorreto"
     },
+    "base64DataTruncated": "[to be translated]:Base64 image data truncated, size",
     "boundary": {
       "default": {
         "devtools": "Abrir o painel de depuração",
@@ -1430,6 +1438,8 @@
     "text": "texto",
     "toolInput": "ferramenta de entrada",
     "toolName": "Nome da ferramenta",
+    "truncated": "[to be translated]:Data truncated, original size",
+    "truncatedBadge": "[to be translated]:Truncated",
     "unknown": "Erro desconhecido",
     "usage": "dosagem",
     "user_message_not_found": "Não foi possível encontrar a mensagem original do usuário",
diff --git a/src/renderer/src/i18n/translate/ro-ro.json b/src/renderer/src/i18n/translate/ro-ro.json
index 9706d6a84d..409f644e6f 100644
--- a/src/renderer/src/i18n/translate/ro-ro.json
+++ b/src/renderer/src/i18n/translate/ro-ro.json
@@ -353,6 +353,13 @@
         "description": "[to be translated]:Create model groups with unique URLs for different provider/model combinations",
         "empty": "[to be translated]:No model groups configured. Click 'Add Group' to create one.",
         "label": "[to be translated]:Model Groups",
+        "mode": {
+          "assistant": "[to be translated]:Assistant Preset",
+          "assistantHint": "[to be translated]:Assistant preset overrides request parameters.",
+          "assistantPlaceholder": "[to be translated]:Select assistant",
+          "label": "[to be translated]:Mode",
+          "model": "[to be translated]:Direct Model"
+        },
         "namePlaceholder": "[to be translated]:Group name"
       },
       "networkAccess": {
@@ -1350,6 +1357,7 @@
     "backup": {
       "file_format": "Eroare format fișier backup"
     },
+    "base64DataTruncated": "[to be translated]:Base64 image data truncated, size",
     "boundary": {
       "default": {
         "devtools": "Deschide panoul de depanare",
@@ -1430,6 +1438,8 @@
     "text": "Text",
     "toolInput": "Intrare instrument",
     "toolName": "Nume instrument",
+    "truncated": "[to be translated]:Data truncated, original size",
+    "truncatedBadge": "[to be translated]:Truncated",
     "unknown": "Eroare necunoscută",
     "usage": "Utilizare",
     "user_message_not_found": "Nu se poate găsi mesajul original al utilizatorului pentru a retrimite",
diff --git a/src/renderer/src/i18n/translate/ru-ru.json b/src/renderer/src/i18n/translate/ru-ru.json
index 235565c202..e938c64f91 100644
--- a/src/renderer/src/i18n/translate/ru-ru.json
+++ b/src/renderer/src/i18n/translate/ru-ru.json
@@ -353,6 +353,13 @@
         "description": "[to be translated]:Create model groups with unique URLs for different provider/model combinations",
         "empty": "[to be translated]:No model groups configured. Click 'Add Group' to create one.",
         "label": "[to be translated]:Model Groups",
+        "mode": {
+          "assistant": "[to be translated]:Assistant Preset",
+          "assistantHint": "[to be translated]:Assistant preset overrides request parameters.",
+          "assistantPlaceholder": "[to be translated]:Select assistant",
+          "label": "[to be translated]:Mode",
+          "model": "[to be translated]:Direct Model"
+        },
         "namePlaceholder": "[to be translated]:Group name"
       },
       "networkAccess": {
@@ -1350,6 +1357,7 @@
     "backup": {
       "file_format": "Ошибка формата файла резервной копии"
     },
+    "base64DataTruncated": "[to be translated]:Base64 image data truncated, size",
     "boundary": {
       "default": {
         "devtools": "Открыть панель отладки",
@@ -1430,6 +1438,8 @@
     "text": "текст",
     "toolInput": "ввод инструмента",
     "toolName": "имя инструмента",
+    "truncated": "[to be translated]:Data truncated, original size",
+    "truncatedBadge": "[to be translated]:Truncated",
     "unknown": "Неизвестная ошибка",
     "usage": "Дозировка",
     "user_message_not_found": "Не удалось найти исходное сообщение пользователя",