fix(config): use deepseek-flash as the default model

.env.example

@@ -381,7 +381,7 @@ RAG_MAX_DISTANCE=0.33
 # DeepSeek
 # DEEPSEEK_API_KEY=<your-api-key>
 DEEPSEEK_BASE_URL=https://api.deepseek.com
-DEEPSEEK_MODEL_NAME=deepseek-v4-flash
+DEEPSEEK_MODEL_NAME=deepseek-flash
 # END GENERATED CONFIG: deepseek
 
 # BEGIN GENERATED CONFIG: gemini
@@ -429,7 +429,7 @@ OPENAI_MODEL_NAME=gpt-5.4-mini
 CODEX_APP_SERVER_CODEX_BIN=""
 # CODEX_APP_SERVER_AUTH_MODE=model_api  # chatgpt | model_api
 CODEX_APP_SERVER_MODEL_PROVIDER=deepseek  # deepseek | openai
-CODEX_APP_SERVER_MODEL=deepseek-v4-flash
+CODEX_APP_SERVER_MODEL=deepseek-flash
 CODEX_APP_SERVER_REASONING_EFFORT=max  # none | minimal | low | medium | high | xhigh | max
 CODEX_APP_SERVER_SERVICE_TIER=""  #  | auto | default | flex | priority | fast | ultrafast
 DIRECT_CODEX_CLI_PATH=codex
@@ -450,7 +450,7 @@ PI_PROVIDER_ENABLED=true
 PI_NODE_PATH=node
 PI_AGENT_DIR=runtime/pi
 PI_MODEL_PROVIDER=deepseek
-PI_MODEL=deepseek-v4-flash
+PI_MODEL=deepseek-flash
 # END GENERATED CONFIG: pi
 
 # BEGIN GENERATED CONFIG: work_routing

README.md

@@ -753,11 +753,11 @@ failure:
 
 | Responsibility | Recommended profile | Current boundary |
 |---|---|---|
-| Main Chat API | DeepSeek-V4-Flash-0731: `DEEPSEEK_BASE_URL=https://api.deepseek.com` and `DEEPSEEK_MODEL_NAME=deepseek-v4-flash` | `deepseek-v4-flash` is the stable API alias currently pointing to the 0731 release; the dated version is not used as the runtime model id. |
+| Main Chat API | DeepSeek Flash: `DEEPSEEK_BASE_URL=https://api.deepseek.com` and `DEEPSEEK_MODEL_NAME=deepseek-flash` | `deepseek-flash` is the recommended API model name; see the [official model-name guidance](https://api-docs.deepseek.com/quick_start/agent_integrations/). Existing explicit model settings are preserved. |
 | Remote speech synthesis | **Fish Audio S2.1**: `TTS_BACKEND=fish_audio`, `FISH_TTS_MODEL=s2.1-pro-free`; Kurisu voice: `FISH_TTS_REFERENCE_ID=b450b19370434173b121446057622e9b` | Bidirectional WebSocket streaming; locally committed sentence chunks use `text → flush`, with incremental audio output. Existing Chat sentence scheduling is preserved. |
 | Multimodal / Vision | Prefer `gemini-3.7-flash`; use `gemini-3.5-flash` as a more conservative compatibility profile | Host-owned visual context performs capture in-process, while image delivery still follows the Main Chat Provider. Independent Gemini Vision API routing is not implemented and does not imply restoring the retired Gemini Live sidecar. |
 | Work execution Provider | Pi for daily tasks; Codex App Server for complex coding | Desktop installation includes the pinned Pi runtime; [configure model credentials](docs/pi-rpc-provider.md). OpenClaw remains optional for explicit selection; Browser retains managed-page operations. |
-| Work execution model | Codex App Server may explicitly select a GPT-5.6-family model or `deepseek-v4-flash` | The execution model belongs to the Work Provider and does not share Main Chat routing or credentials. |
+| Work execution model | Codex App Server may explicitly select a GPT-5.6-family model or `deepseek-flash` | The execution model belongs to the Work Provider and does not share Main Chat routing or credentials. |
 | AUIP runtime action decisions | `AUIP_ACTION_PROVIDER=openai`, `AUIP_ACTION_MODEL=gpt-5.6-terra`, `AUIP_ACTION_REASONING_EFFORT=low`, and `AUIP_ACTION_SERVICE_TIER=fast` | This model decides AppSession actions and participation; it is not the execution Provider that authors an AUIP Artifact. `fast` requires availability for the API project. |
 
 ### Recommended remote TTS: Fish Audio + Kurisu

config/catalog/models/deepseek.json

@@ -42,7 +42,7 @@
         "en-US": "Independent from the Codex Work Provider model.",
         "zh-CN": "独立于 Codex Work Provider 使用的模型。"
       },
-      "default": "deepseek-v4-flash",
+      "default": "deepseek-flash",
       "example_active": true
     }
   },

config/catalog/providers/codex.json

@@ -125,7 +125,7 @@
         "zh-CN": "默认继承所选 Models 连接中的模型。"
       },
       "computed_default": true,
-      "example": "deepseek-v4-flash",
+      "example": "deepseek-flash",
       "visible_when": {
         "CODEX_PROVIDER_TRANSPORT": [
           "app_server"

config/catalog/providers/pi.json

@@ -57,7 +57,7 @@
         "zh-CN": "Pi 模型"
       },
       "computed_default": true,
-      "example": "deepseek-v4-flash",
+      "example": "deepseek-flash",
       "example_active": true
     }
   },

electron/src/renderer/components/SettingsPage.tsx

@@ -1073,7 +1073,7 @@ export default function SettingsPage({ send, subscribe, connected, reconnectBack
   const mainModelLabel = modelProviderLabels[mainModelProvider] || mainModelProvider
   const imageModel = mainModelProvider === 'deepseek' || mainModelProvider === 'hybrid2'
     ? String(modelConnections.find(group => group.id === 'deepseek')?.fields
-      ?.find(field => field.key === 'DEEPSEEK_MODEL_NAME')?.value || 'deepseek-v4-flash')
+      ?.find(field => field.key === 'DEEPSEEK_MODEL_NAME')?.value || 'deepseek-flash')
     : ''
   const visionModelReady = mainModelReady && chatProviderSupportsImages(mainModelProvider, imageModel)
   const configuredTranslationModels = modelConnections

electron/src/renderer/components/sceneCapabilityProjection.ts

@@ -323,7 +323,7 @@ function sceneVisualCapability(
   const modelConnections = groups(settings.model_connections)
   const requiredConnections = chatProfileConnections(provider)
   const imageModel = provider === 'deepseek' || provider === 'hybrid2'
-    ? fieldValue(findGroup(modelConnections, 'deepseek'), 'DEEPSEEK_MODEL_NAME', 'deepseek-v4-flash')
+    ? fieldValue(findGroup(modelConnections, 'deepseek'), 'DEEPSEEK_MODEL_NAME', 'deepseek-flash')
     : ''
   const supported = chatProviderSupportsImages(provider, imageModel)
   const modelConfigured = requiredConnections.length > 0

server/vn_tts_bridge.py

@@ -767,7 +767,7 @@ async def _translate_ja_to_zh(japanese_text: str) -> str:
     else:
         api_key = getattr(settings, "DEEPSEEK_API_KEY", "")
         base_url = os.environ.get("VN_SUBTITLE_TRANSLATE_BASE_URL") or getattr(settings, "DEEPSEEK_BASE_URL", "")
-        model = read_catalog_value(EnvironmentReader(os.environ), "VN_SUBTITLE_TRANSLATE_MODEL") or getattr(settings, "DEEPSEEK_MODEL_NAME", "deepseek-v4-flash")
+        model = read_catalog_value(EnvironmentReader(os.environ), "VN_SUBTITLE_TRANSLATE_MODEL") or getattr(settings, "DEEPSEEK_MODEL_NAME", "deepseek-flash")
     if not api_key:
         raise RuntimeError(f"{provider} API key is not configured")
 
@@ -814,7 +814,7 @@ async def _stream_translate_zh_to_ja(chinese_text: str) -> AsyncIterator[str]:
     else:
         api_key = getattr(settings, "DEEPSEEK_API_KEY", "")
         base_url = os.environ.get("VN_TTS_TRANSLATE_BASE_URL") or getattr(settings, "DEEPSEEK_BASE_URL", "")
-        model = read_catalog_value(EnvironmentReader(os.environ), "VN_TTS_TRANSLATE_MODEL") or getattr(settings, "DEEPSEEK_MODEL_NAME", "deepseek-v4-flash")
+        model = read_catalog_value(EnvironmentReader(os.environ), "VN_TTS_TRANSLATE_MODEL") or getattr(settings, "DEEPSEEK_MODEL_NAME", "deepseek-flash")
     if not api_key:
         raise RuntimeError(f"{provider} API key is not configured")
 

server/wallpaper_subtitle_translator.py

@@ -152,7 +152,7 @@ def _config_for_provider(base: SubtitleTranslatorConfig, provider: str) -> Subti
         base_url = base_url_override or "https://generativelanguage.googleapis.com/v1beta"
     else:
         provider = "deepseek"
-        model = model_override or getattr(settings, "DEEPSEEK_MODEL_NAME", "deepseek-v4-flash")
+        model = model_override or getattr(settings, "DEEPSEEK_MODEL_NAME", "deepseek-flash")
         base_url = base_url_override or getattr(settings, "DEEPSEEK_BASE_URL", "")
 
     return SubtitleTranslatorConfig(

tests/test_config_catalog.py

@@ -116,6 +116,14 @@ def test_catalog_defaults_remain_registered_in_the_environment_reader() -> None:
             }.get(field["type"], "str"))
 
 
+@pytest.mark.parametrize("model", [None, "deepseek-v4-flash", "deepseek-v4-pro", "custom-model", ""])
+def test_deepseek_default_preserves_explicit_model_selection(model) -> None:
+    reader = EnvironmentReader({} if model is None else {"DEEPSEEK_MODEL_NAME": model})
+    assert catalog.read_catalog_value(reader, "DEEPSEEK_MODEL_NAME") == (
+        "deepseek-flash" if model is None else model
+    )
+
+
 def test_catalog_preserves_explicit_empty_values_and_secret_normalization() -> None:
     reader = EnvironmentReader({
         "FISH_TTS_MODEL": "", "FISH_TTS_LATENCY": "low",