fix(cli): accept the qwen-audio Token Plan ASR family in voice transport resolution

docs/users/configuration/settings.md

@@ -91,7 +91,7 @@ Settings are organized into categories. Most settings should be placed within th
 | `general.voice.enabled`                    | boolean | Enable voice dictation in the prompt input. Also toggleable with the `/voice` command. Requires a transcription model (`voiceModel`) to be configured.                                                                                                                                                                                                                                                                                                                | `false`     |
 | `general.voice.mode`                       | enum    | How push-to-talk behaves: `"hold"` to talk while the key is held, or `"tap"` to start and tap (or pause) to stop and submit.                                                                                                                                                                                                                                                                                                                                          | `"hold"`    |
 | `general.voice.language`                   | string  | Preferred spoken language for voice transcription (e.g. `"english"`, `"chinese"`). Leave empty to auto-detect.                                                                                                                                                                                                                                                                                                                                                        | `""`        |
-| `general.voice.keytermsFile`               | string  | Path to a custom keyterms file (one term per line, `#` for comments) that biases voice transcription toward domain-specific terms. Relative paths resolve from the workspace root; defaults to `.qwen/voice-keyterms.txt` when present. Read only in trusted workspaces. Only applies to Qwen ASR models (`qwen3-asr-*`).                                                                                                                                             | `""`        |
+| `general.voice.keytermsFile`               | string  | Path to a custom keyterms file (one term per line, `#` for comments) that biases voice transcription toward domain-specific terms. Relative paths resolve from the workspace root; defaults to `.qwen/voice-keyterms.txt` when present. Read only in trusted workspaces. Only applies to Qwen ASR models (`qwen3-asr-*` and the `qwen-audio-*` ASR family).                                                                                                           | `""`        |
 | `general.voice.refineTranscript`           | boolean | Clean up voice transcripts with the fast model before inserting them — removes filler words and fixes recognition errors while preserving meaning. Falls back to the raw transcript on failure, and is skipped when no fast model is configured.                                                                                                                                                                                                                      | `true`      |
 | `general.cleanupPeriodDays`                | number  | Days to retain `~/.qwen/file-history/` session backups used by `/rewind`. Backups older than this are removed by a background pass that runs at most once per day. `0` = minimum retention (~1 hour): keeps sessions touched in the last hour plus the currently active one. Changes take effect after restart.                                                                                                                                                       | `30`        |
 | `general.language`                         | enum    | Language for the user interface. Use `"auto"` to detect from system settings, or a language code (e.g. `"zh-CN"`, `"fr"`). Custom codes can be added by placing JS locale files in `~/.qwen/locales/`. See [i18n](../features/language). Requires restart.                                                                                                                                                                                                            | `"auto"`    |

packages/cli/src/config/settingsSchema.ts

@@ -506,7 +506,7 @@ const SETTINGS_SCHEMA = {
             requiresRestart: false,
             default: '',
             description:
-              'Path to a custom keyterms file (one term per line, "#" for comments) that biases voice transcription toward domain-specific terms. Relative paths resolve from the workspace root; defaults to ".qwen/voice-keyterms.txt" when present. The file contents are sent to the ASR provider and it is read only in trusted workspaces. Only applies to Qwen ASR models (qwen3-asr-*).',
+              'Path to a custom keyterms file (one term per line, "#" for comments) that biases voice transcription toward domain-specific terms. Relative paths resolve from the workspace root; defaults to ".qwen/voice-keyterms.txt" when present. The file contents are sent to the ASR provider and it is read only in trusted workspaces. Only applies to Qwen ASR models (qwen3-asr-* and the qwen-audio-* ASR family).',
             showInDialog: false,
           },
           refineTranscript: {

packages/cli/src/services/voice-model.ts

@@ -13,7 +13,15 @@ export type VoiceTransport =
   | 'dashscope-task-realtime'
   | 'unsupported';
 
-/** Map a model id to the ASR transport it uses, or 'unsupported'. */
+/**
+ * Map a model id to the ASR transport it uses, or 'unsupported'.
+ *
+ * The `qwen-audio-<version>` family (Model Studio Token Plan, #10932) follows
+ * the same split as `qwen3-asr-flash`: `*-realtime` and `*-asr-flash-streaming`
+ * ids speak the OpenAI realtime WebSocket dialect, while bare and date-suffixed
+ * `*-asr-flash` ids use the batch chat/completions shape. Other ids in the
+ * family (filetrans, tts) are not dictation transports and stay unsupported.
+ */
 export function resolveVoiceTransport(model: string): VoiceTransport {
   const id = model.toLowerCase();
   if (/^qwen3-asr-flash-realtime(?:-|$)/.test(id)) {
@@ -22,6 +30,16 @@ export function resolveVoiceTransport(model: string): VoiceTransport {
   if (/^qwen3-asr-flash(?:-\d{4}-\d{2}-\d{2})?$/.test(id)) {
     return 'qwen-asr-chat';
   }
+  if (
+    /^qwen-audio-[\d.]+-(?:asr-flash-(?:realtime|streaming)|realtime)(?:-|$)/.test(
+      id,
+    )
+  ) {
+    return 'qwen-asr-realtime';
+  }
+  if (/^qwen-audio-[\d.]+-asr-flash(?:-\d{4}-\d{2}-\d{2})?$/.test(id)) {
+    return 'qwen-asr-chat';
+  }
   if (/^(fun-asr|paraformer).*realtime(?:-|$)/.test(id)) {
     return 'dashscope-task-realtime';
   }

packages/cli/src/ui/voice/voice-model.test.ts

@@ -9,6 +9,7 @@ import { AuthType, type AvailableModel } from '@qwen-code/qwen-code-core';
 import {
   isSelectableVoiceModel,
   isTranscribableVoiceModel,
+  resolveVoiceTransport,
 } from './voice-model.js';
 
 function model(overrides: Partial<AvailableModel>): AvailableModel {
@@ -51,4 +52,68 @@ describe('voice model guards', () => {
       isSelectableVoiceModel(model({ id: 'qwen3-asr-flash-filetrans' })),
     ).toBe(false);
   });
+
+  it('isSelectableVoiceModel accepts the qwen-audio Token Plan ASR family', () => {
+    expect(
+      isSelectableVoiceModel(model({ id: 'qwen-audio-3.0-asr-flash' })),
+    ).toBe(true);
+    expect(
+      isSelectableVoiceModel(model({ id: 'qwen-audio-3.0-realtime-plus' })),
+    ).toBe(true);
+  });
+});
+
+describe('resolveVoiceTransport', () => {
+  it('routes the qwen-audio Token Plan ASR family', () => {
+    // Bare and date-suffixed asr ids use the batch chat shape (#10932).
+    expect(resolveVoiceTransport('qwen-audio-3.0-asr-flash')).toBe(
+      'qwen-asr-chat',
+    );
+    expect(resolveVoiceTransport('qwen-audio-3.0-asr-flash-2026-09-01')).toBe(
+      'qwen-asr-chat',
+    );
+    // Streaming variants speak the OpenAI realtime WebSocket dialect.
+    expect(resolveVoiceTransport('qwen-audio-3.0-asr-flash-streaming')).toBe(
+      'qwen-asr-realtime',
+    );
+    expect(resolveVoiceTransport('qwen-audio-3.0-asr-flash-realtime')).toBe(
+      'qwen-asr-realtime',
+    );
+    expect(resolveVoiceTransport('qwen-audio-3.0-realtime-plus')).toBe(
+      'qwen-asr-realtime',
+    );
+    expect(resolveVoiceTransport('Qwen-Audio-3.0-ASR-Flash')).toBe(
+      'qwen-asr-chat',
+    );
+  });
+
+  it('keeps routing the legacy qwen3-asr and dashscope ids', () => {
+    expect(resolveVoiceTransport('qwen3-asr-flash')).toBe('qwen-asr-chat');
+    expect(resolveVoiceTransport('qwen3-asr-flash-2025-08-20')).toBe(
+      'qwen-asr-chat',
+    );
+    expect(resolveVoiceTransport('qwen3-asr-flash-realtime')).toBe(
+      'qwen-asr-realtime',
+    );
+    expect(resolveVoiceTransport('fun-asr-realtime')).toBe(
+      'dashscope-task-realtime',
+    );
+    expect(resolveVoiceTransport('paraformer-realtime-v2')).toBe(
+      'dashscope-task-realtime',
+    );
+  });
+
+  it('rejects non-ASR ids, including filetrans and tts in the family', () => {
+    expect(resolveVoiceTransport('gpt-4o')).toBe('unsupported');
+    expect(resolveVoiceTransport('custom:asr')).toBe('unsupported');
+    expect(resolveVoiceTransport('qwen3-asr-flash-filetrans')).toBe(
+      'unsupported',
+    );
+    expect(resolveVoiceTransport('qwen-audio-3.0-asr-flash-filetrans')).toBe(
+      'unsupported',
+    );
+    expect(resolveVoiceTransport('qwen-audio-3.0-tts-plus')).toBe(
+      'unsupported',
+    );
+  });
 });

packages/web-shell/client/voice/voiceModels.test.ts

@@ -0,0 +1,76 @@
+/**
+ * @license
+ * Copyright 2025 Qwen
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+import { describe, expect, it } from 'vitest';
+import { extractVoiceModels, isVoiceModelId } from './voiceModels';
+
+describe('isVoiceModelId', () => {
+  it('accepts the qwen-audio Token Plan ASR family (#10932)', () => {
+    expect(isVoiceModelId('qwen-audio-3.0-asr-flash')).toBe(true);
+    expect(isVoiceModelId('qwen-audio-3.0-asr-flash-2026-09-01')).toBe(true);
+    expect(isVoiceModelId('qwen-audio-3.0-asr-flash-streaming')).toBe(true);
+    expect(isVoiceModelId('qwen-audio-3.0-asr-flash-realtime')).toBe(true);
+    expect(isVoiceModelId('qwen-audio-3.0-realtime-plus')).toBe(true);
+    expect(isVoiceModelId('Qwen-Audio-3.0-ASR-Flash')).toBe(true);
+  });
+
+  it('keeps accepting the legacy ASR ids', () => {
+    expect(isVoiceModelId('qwen3-asr-flash')).toBe(true);
+    expect(isVoiceModelId('qwen3-asr-flash-2025-08-20')).toBe(true);
+    expect(isVoiceModelId('qwen3-asr-flash-realtime')).toBe(true);
+    expect(isVoiceModelId('fun-asr-realtime')).toBe(true);
+    expect(isVoiceModelId('paraformer-realtime-v2')).toBe(true);
+  });
+
+  it('rejects non-ASR ids, including filetrans and tts in the family', () => {
+    expect(isVoiceModelId('gpt-4o')).toBe(false);
+    expect(isVoiceModelId('qwen3-asr-flash-filetrans')).toBe(false);
+    expect(isVoiceModelId('qwen-audio-3.0-asr-flash-filetrans')).toBe(false);
+    expect(isVoiceModelId('qwen-audio-3.0-tts-plus')).toBe(false);
+  });
+});
+
+describe('extractVoiceModels', () => {
+  it('lists qwen-audio ASR ids from a providers status and dedupes', () => {
+    const status = {
+      providers: [
+        {
+          authType: 'USE_OPENAI',
+          models: [
+            {
+              baseModelId: 'qwen-audio-3.0-asr-flash',
+              name: 'Qwen Audio ASR',
+              baseUrl: 'https://token-plan.example/compatible-mode/v1',
+              contextLimit: 8000,
+            },
+            {
+              // Registered with an auth suffix; the base id is the voice id.
+              baseModelId: 'qwen-audio-3.0-asr-flash',
+              name: 'Qwen Audio ASR (duplicate provider)',
+            },
+            { baseModelId: 'qwen-max', name: 'Qwen Max' },
+            { baseModelId: 'qwen-audio-3.0-asr-flash-filetrans' },
+            { isRuntime: true, baseModelId: 'qwen3-asr-flash-realtime' },
+          ],
+        },
+      ],
+    };
+    const options = extractVoiceModels(status);
+    expect(options).toHaveLength(1);
+    expect(options[0]).toEqual({
+      id: 'qwen-audio-3.0-asr-flash',
+      label: 'Qwen Audio ASR',
+      authType: 'USE_OPENAI',
+      baseUrl: 'https://token-plan.example/compatible-mode/v1',
+      contextWindow: 8000,
+      modalities: { audio: true },
+    });
+  });
+
+  it('returns an empty list for an undefined status', () => {
+    expect(extractVoiceModels(undefined)).toEqual([]);
+  });
+});

packages/web-shell/client/voice/voiceModels.ts

@@ -25,6 +25,10 @@ export function isVoiceModelId(id: string): boolean {
   return (
     /^qwen3-asr-flash-realtime(?:-|$)/.test(s) ||
     /^qwen3-asr-flash(?:-\d{4}-\d{2}-\d{2})?$/.test(s) ||
+    /^qwen-audio-[\d.]+-(?:asr-flash-(?:realtime|streaming)|realtime)(?:-|$)/.test(
+      s,
+    ) ||
+    /^qwen-audio-[\d.]+-asr-flash(?:-\d{4}-\d{2}-\d{2})?$/.test(s) ||
     /^(fun-asr|paraformer).*realtime(?:-|$)/.test(s)
   );
 }