fix(web): preserve explicit model providers

web/backend/api/model_status.go

@@ -87,7 +87,7 @@ func hasModelConfiguration(m *config.ModelConfig) bool {
 	apiKey := strings.TrimSpace(m.APIKey())
 
 	if authMethod == "oauth" || authMethod == "token" {
-		if provider, ok := oauthProviderForModel(m.Model); ok {
+		if provider, ok := oauthProviderForModel(m); ok {
 			cred, err := oauthGetCredential(provider)
 			if err != nil || cred == nil {
 				return false
@@ -123,7 +123,7 @@ func requiresRuntimeProbe(m *config.ModelConfig) bool {
 		return true
 	}
 
-	protocol := modelProtocol(m.Model)
+	protocol := modelProtocol(m)
 
 	switch protocol {
 	case "claude-cli", "claudecli", "codex-cli", "codexcli", "github-copilot", "copilot":
@@ -172,7 +172,7 @@ func (s *modelProbeCacheState) probe(cacheKey string, probeFunc func() bool) boo
 
 func runLocalModelProbe(m *config.ModelConfig) bool {
 	apiBase := modelProbeAPIBase(m)
-	protocol, modelID := splitModel(m.Model)
+	protocol, modelID := splitModel(m)
 	switch protocol {
 	case "ollama":
 		return probeOllamaModelFunc(apiBase, modelID)
@@ -191,7 +191,7 @@ func runLocalModelProbe(m *config.ModelConfig) bool {
 }
 
 func modelProbeCacheKey(m *config.ModelConfig) string {
-	protocol, modelID := splitModel(m.Model)
+	protocol, modelID := splitModel(m)
 
 	apiBaseRaw := modelProbeAPIBase(m)
 	apiBase := strings.ToLower(strings.TrimRight(strings.TrimSpace(apiBaseRaw), "/"))
@@ -384,7 +384,7 @@ func modelProbeAPIBase(m *config.ModelConfig) string {
 		return normalizeModelProbeAPIBase(apiBase)
 	}
 
-	protocol := modelProtocol(m.Model)
+	protocol := modelProtocol(m)
 	if providers.IsEmptyAPIKeyAllowedForProtocol(protocol) {
 		return providers.DefaultAPIBaseForProtocol(protocol)
 	}
@@ -419,8 +419,8 @@ func normalizeModelProbeAPIBase(raw string) string {
 	return u.String()
 }
 
-func oauthProviderForModel(model string) (string, bool) {
+func oauthProviderForModel(m *config.ModelConfig) (string, bool) {
-	switch modelProtocol(model) {
+	switch modelProtocol(m) {
 	case "openai":
 		return oauthProviderOpenAI, true
 	case "anthropic":
@@ -432,18 +432,14 @@ func oauthProviderForModel(model string) (string, bool) {
 	}
 }
 
-func modelProtocol(model string) string {
+func modelProtocol(m *config.ModelConfig) string {
-	protocol, _ := splitModel(model)
+	protocol, _ := splitModel(m)
 	return protocol
 }
 
-func splitModel(model string) (protocol, modelID string) {
+func splitModel(m *config.ModelConfig) (protocol, modelID string) {
-	model = strings.ToLower(strings.TrimSpace(model))
+	protocol, modelID = providers.ExtractProtocol(m)
-	protocol, _, found := strings.Cut(model, "/")
+	return strings.ToLower(strings.TrimSpace(protocol)), strings.ToLower(strings.TrimSpace(modelID))
-	if !found {
-		return "openai", model
-	}
-	return protocol, strings.TrimSpace(model[strings.Index(model, "/")+1:])
 }
 
 func hasLocalAPIBase(raw string) bool {

web/backend/api/models.go

@@ -10,6 +10,7 @@ import (
 
 	"github.com/sipeed/picoclaw/pkg/config"
 	"github.com/sipeed/picoclaw/pkg/logger"
+	"github.com/sipeed/picoclaw/pkg/providers"
 )
 
 // registerModelRoutes binds model list management endpoints to the ServeMux.
@@ -26,6 +27,7 @@ func (h *Handler) registerModelRoutes(mux *http.ServeMux) {
 type modelResponse struct {
 	Index      int    `json:"index"`
 	ModelName  string `json:"model_name"`
+	Provider   string `json:"provider,omitempty"`
 	Model      string `json:"model"`
 	APIBase    string `json:"api_base,omitempty"`
 	APIKey     string `json:"api_key"`
@@ -73,10 +75,12 @@ func (h *Handler) handleListModels(w http.ResponseWriter, r *http.Request) {
 
 	models := make([]modelResponse, 0, len(cfg.ModelList))
 	for i, m := range cfg.ModelList {
+		provider, modelID := providers.ExtractProtocol(m)
 		models = append(models, modelResponse{
 			Index:          i,
 			ModelName:      m.ModelName,
-			Model:          m.Model,
+			Provider:       provider,
+			Model:          modelID,
 			APIBase:        m.APIBase,
 			APIKey:         maskAPIKey(m.APIKey()),
 			Proxy:          m.Proxy,
@@ -176,6 +180,12 @@ func (h *Handler) handleUpdateModel(w http.ResponseWriter, r *http.Request) {
 	}
 	defer r.Body.Close()
 
+	var rawFields map[string]json.RawMessage
+	if err = json.Unmarshal(body, &rawFields); err != nil {
+		http.Error(w, fmt.Sprintf("Invalid JSON: %v", err), http.StatusBadRequest)
+		return
+	}
+
 	type custom struct {
 		config.ModelConfig
 		APIKey string `json:"api_key"`
@@ -226,6 +236,12 @@ func (h *Handler) handleUpdateModel(w http.ResponseWriter, r *http.Request) {
 	} else if len(mc.CustomHeaders) == 0 {
 		mc.CustomHeaders = nil
 	}
+	// Preserve the existing Provider when the caller omits it. This keeps the
+	// update API backward-compatible for clients that haven't started sending
+	// the new field yet, while still allowing explicit clearing via "".
+	if _, ok := rawFields["provider"]; !ok {
+		mc.Provider = cfg.ModelList[idx].Provider
+	}
 
 	cfg.ModelList[idx] = &mc.ModelConfig
 

web/backend/api/models_test.go

@@ -392,6 +392,49 @@ func TestHandleListModels_StatusMarksUnreachableLocalModel(t *testing.T) {
 	}
 }
 
+func TestHandleListModels_RuntimeProbeUsesExplicitProviderField(t *testing.T) {
+	configPath, cleanup := setupOAuthTestEnv(t)
+	defer cleanup()
+	resetOAuthHooks(t)
+	resetModelProbeHooks(t)
+
+	var gotProbe string
+	probeOpenAICompatibleModelFunc = func(apiBase, modelID, apiKey string) bool {
+		gotProbe = apiBase + "|" + modelID + "|" + apiKey
+		return true
+	}
+
+	cfg, err := config.LoadConfig(configPath)
+	if err != nil {
+		t.Fatalf("LoadConfig() error = %v", err)
+	}
+	cfg.ModelList = []*config.ModelConfig{{
+		ModelName: "vllm-local",
+		Provider:  "vllm",
+		Model:     "custom-model",
+		APIBase:   "http://127.0.0.1:8000/v1",
+	}}
+	if err := config.SaveConfig(configPath, cfg); err != nil {
+		t.Fatalf("SaveConfig() error = %v", err)
+	}
+
+	h := NewHandler(configPath)
+	mux := http.NewServeMux()
+	h.RegisterRoutes(mux)
+
+	rec := httptest.NewRecorder()
+	req := httptest.NewRequest(http.MethodGet, "/api/models", nil)
+	mux.ServeHTTP(rec, req)
+
+	if rec.Code != http.StatusOK {
+		t.Fatalf("status = %d, want %d, body=%s", rec.Code, http.StatusOK, rec.Body.String())
+	}
+
+	if gotProbe != "http://127.0.0.1:8000/v1|custom-model|" {
+		t.Fatalf("probe = %q, want %q", gotProbe, "http://127.0.0.1:8000/v1|custom-model|")
+	}
+}
+
 func TestHandleAddModel_PersistsAPIKey(t *testing.T) {
 	configPath, cleanup := setupOAuthTestEnv(t)
 	defer cleanup()
@@ -430,6 +473,41 @@ func TestHandleAddModel_PersistsAPIKey(t *testing.T) {
 	}
 }
 
+func TestHandleAddModel_PersistsProvider(t *testing.T) {
+	configPath, cleanup := setupOAuthTestEnv(t)
+	defer cleanup()
+
+	h := NewHandler(configPath)
+	mux := http.NewServeMux()
+	h.RegisterRoutes(mux)
+
+	rec := httptest.NewRecorder()
+	req := httptest.NewRequest(http.MethodPost, "/api/models", bytes.NewBufferString(`{
+		"model_name":"nvidia-glm",
+		"provider":"nvidia",
+		"model":"z-ai/glm-5.1",
+		"api_key":"nv-key"
+	}`))
+	req.Header.Set("Content-Type", "application/json")
+	mux.ServeHTTP(rec, req)
+
+	if rec.Code != http.StatusOK {
+		t.Fatalf("status = %d, want %d, body=%s", rec.Code, http.StatusOK, rec.Body.String())
+	}
+
+	cfg, err := config.LoadConfig(configPath)
+	if err != nil {
+		t.Fatalf("LoadConfig() error = %v", err)
+	}
+	added := cfg.ModelList[len(cfg.ModelList)-1]
+	if added.Provider != "nvidia" {
+		t.Fatalf("provider = %q, want %q", added.Provider, "nvidia")
+	}
+	if added.Model != "z-ai/glm-5.1" {
+		t.Fatalf("model = %q, want %q", added.Model, "z-ai/glm-5.1")
+	}
+}
+
 func TestHandleAddModel_PersistsCustomHeaders(t *testing.T) {
 	configPath, cleanup := setupOAuthTestEnv(t)
 	defer cleanup()
@@ -536,6 +614,206 @@ func TestHandleUpdateModel_CustomHeadersPreserveAndClear(t *testing.T) {
 	}
 }
 
+func TestHandleUpdateModel_PersistsProvider(t *testing.T) {
+	configPath, cleanup := setupOAuthTestEnv(t)
+	defer cleanup()
+
+	cfg, err := config.LoadConfig(configPath)
+	if err != nil {
+		t.Fatalf("LoadConfig() error = %v", err)
+	}
+	cfg.ModelList = []*config.ModelConfig{{
+		ModelName: "editable",
+		Model:     "gpt-4o",
+		Provider:  "openai",
+	}}
+	if err := config.SaveConfig(configPath, cfg); err != nil {
+		t.Fatalf("SaveConfig() error = %v", err)
+	}
+
+	h := NewHandler(configPath)
+	mux := http.NewServeMux()
+	h.RegisterRoutes(mux)
+
+	rec := httptest.NewRecorder()
+	req := httptest.NewRequest(http.MethodPut, "/api/models/0", bytes.NewBufferString(`{
+		"model_name":"editable",
+		"provider":"openrouter",
+		"model":"openai/gpt-4o"
+	}`))
+	req.Header.Set("Content-Type", "application/json")
+	mux.ServeHTTP(rec, req)
+
+	if rec.Code != http.StatusOK {
+		t.Fatalf("status = %d, want %d, body=%s", rec.Code, http.StatusOK, rec.Body.String())
+	}
+
+	updated, err := config.LoadConfig(configPath)
+	if err != nil {
+		t.Fatalf("LoadConfig() error = %v", err)
+	}
+	if got := updated.ModelList[0].Provider; got != "openrouter" {
+		t.Fatalf("provider = %q, want %q", got, "openrouter")
+	}
+}
+
+func TestHandleUpdateModel_PreservesProviderWhenOmitted(t *testing.T) {
+	configPath, cleanup := setupOAuthTestEnv(t)
+	defer cleanup()
+
+	cfg, err := config.LoadConfig(configPath)
+	if err != nil {
+		t.Fatalf("LoadConfig() error = %v", err)
+	}
+	cfg.ModelList = []*config.ModelConfig{{
+		ModelName: "editable",
+		Model:     "z-ai/glm-5.1",
+		Provider:  "nvidia",
+	}}
+	if err := config.SaveConfig(configPath, cfg); err != nil {
+		t.Fatalf("SaveConfig() error = %v", err)
+	}
+
+	h := NewHandler(configPath)
+	mux := http.NewServeMux()
+	h.RegisterRoutes(mux)
+
+	rec := httptest.NewRecorder()
+	req := httptest.NewRequest(http.MethodPut, "/api/models/0", bytes.NewBufferString(`{
+		"model_name":"editable",
+		"model":"z-ai/glm-5.1"
+	}`))
+	req.Header.Set("Content-Type", "application/json")
+	mux.ServeHTTP(rec, req)
+
+	if rec.Code != http.StatusOK {
+		t.Fatalf("status = %d, want %d, body=%s", rec.Code, http.StatusOK, rec.Body.String())
+	}
+
+	updated, err := config.LoadConfig(configPath)
+	if err != nil {
+		t.Fatalf("LoadConfig() error = %v", err)
+	}
+	if got := updated.ModelList[0].Provider; got != "nvidia" {
+		t.Fatalf("provider = %q, want %q", got, "nvidia")
+	}
+}
+
+func TestHandleListModels_ReturnsProviderField(t *testing.T) {
+	configPath, cleanup := setupOAuthTestEnv(t)
+	defer cleanup()
+
+	cfg, err := config.LoadConfig(configPath)
+	if err != nil {
+		t.Fatalf("LoadConfig() error = %v", err)
+	}
+	cfg.ModelList = []*config.ModelConfig{{
+		ModelName: "nvidia-glm",
+		Provider:  "nvidia",
+		Model:     "z-ai/glm-5.1",
+		APIKeys:   config.SimpleSecureStrings("nv-key"),
+	}}
+	if err := config.SaveConfig(configPath, cfg); err != nil {
+		t.Fatalf("SaveConfig() error = %v", err)
+	}
+
+	h := NewHandler(configPath)
+	mux := http.NewServeMux()
+	h.RegisterRoutes(mux)
+
+	rec := httptest.NewRecorder()
+	req := httptest.NewRequest(http.MethodGet, "/api/models", nil)
+	mux.ServeHTTP(rec, req)
+
+	if rec.Code != http.StatusOK {
+		t.Fatalf("status = %d, want %d, body=%s", rec.Code, http.StatusOK, rec.Body.String())
+	}
+
+	var resp struct {
+		Models []modelResponse `json:"models"`
+	}
+	if err := json.Unmarshal(rec.Body.Bytes(), &resp); err != nil {
+		t.Fatalf("Unmarshal() error = %v", err)
+	}
+	if len(resp.Models) != 1 {
+		t.Fatalf("len(models) = %d, want 1", len(resp.Models))
+	}
+	if got := resp.Models[0].Provider; got != "nvidia" {
+		t.Fatalf("provider = %q, want %q", got, "nvidia")
+	}
+}
+
+func TestHandleListModels_ReturnsEffectiveProviderField(t *testing.T) {
+	configPath, cleanup := setupOAuthTestEnv(t)
+	defer cleanup()
+
+	cfg, err := config.LoadConfig(configPath)
+	if err != nil {
+		t.Fatalf("LoadConfig() error = %v", err)
+	}
+	cfg.ModelList = []*config.ModelConfig{
+		{
+			ModelName: "plain-openai",
+			Model:     "gpt-4o",
+		},
+		{
+			ModelName: "explicit-google",
+			Provider:  "google",
+			Model:     "gemini-2.5-pro",
+		},
+		{
+			ModelName: "explicit-qwen-intl",
+			Provider:  "qwen-international",
+			Model:     "qwen3-coder-plus",
+		},
+	}
+	if err := config.SaveConfig(configPath, cfg); err != nil {
+		t.Fatalf("SaveConfig() error = %v", err)
+	}
+
+	h := NewHandler(configPath)
+	mux := http.NewServeMux()
+	h.RegisterRoutes(mux)
+
+	rec := httptest.NewRecorder()
+	req := httptest.NewRequest(http.MethodGet, "/api/models", nil)
+	mux.ServeHTTP(rec, req)
+
+	if rec.Code != http.StatusOK {
+		t.Fatalf("status = %d, want %d, body=%s", rec.Code, http.StatusOK, rec.Body.String())
+	}
+
+	var resp struct {
+		Models []modelResponse `json:"models"`
+	}
+	if err := json.Unmarshal(rec.Body.Bytes(), &resp); err != nil {
+		t.Fatalf("Unmarshal() error = %v", err)
+	}
+
+	if len(resp.Models) != 3 {
+		t.Fatalf("len(models) = %d, want 3", len(resp.Models))
+	}
+
+	if got := resp.Models[0].Provider; got != "openai" {
+		t.Fatalf("provider[0] = %q, want %q", got, "openai")
+	}
+	if got := resp.Models[0].Model; got != "gpt-4o" {
+		t.Fatalf("model[0] = %q, want %q", got, "gpt-4o")
+	}
+	if got := resp.Models[1].Provider; got != "gemini" {
+		t.Fatalf("provider[1] = %q, want %q", got, "gemini")
+	}
+	if got := resp.Models[1].Model; got != "gemini-2.5-pro" {
+		t.Fatalf("model[1] = %q, want %q", got, "gemini-2.5-pro")
+	}
+	if got := resp.Models[2].Provider; got != "qwen-intl" {
+		t.Fatalf("provider[2] = %q, want %q", got, "qwen-intl")
+	}
+	if got := resp.Models[2].Model; got != "qwen3-coder-plus" {
+		t.Fatalf("model[2] = %q, want %q", got, "qwen3-coder-plus")
+	}
+}
+
 // TestHandleSetDefaultModel_RejectsNonexistentModel tests that setting a non-existent
 // model as default returns 404. This covers the case where virtual models (which are
 // filtered by SaveConfig) cannot be set as default.

web/backend/api/oauth.go

@@ -746,7 +746,7 @@ func (h *Handler) syncProviderAuthMethod(provider, authMethod string) error {
 
 	found := false
 	for i := range cfg.ModelList {
-		if modelBelongsToProvider(provider, cfg.ModelList[i].Model) {
+		if modelBelongsToProvider(provider, cfg.ModelList[i]) {
 			cfg.ModelList[i].AuthMethod = authMethod
 			found = true
 		}
@@ -759,18 +759,15 @@ func (h *Handler) syncProviderAuthMethod(provider, authMethod string) error {
 	return oauthSaveConfig(h.configPath, cfg)
 }
 
-func modelBelongsToProvider(provider, model string) bool {
+func modelBelongsToProvider(provider string, modelCfg *config.ModelConfig) bool {
-	lower := strings.ToLower(strings.TrimSpace(model))
+	protocol, _ := providers.ExtractProtocol(modelCfg)
 	switch provider {
 	case oauthProviderOpenAI:
-		return lower == "openai" || strings.HasPrefix(lower, "openai/")
+		return protocol == "openai"
 	case oauthProviderAnthropic:
-		return lower == "anthropic" || strings.HasPrefix(lower, "anthropic/")
+		return protocol == "anthropic"
 	case oauthProviderGoogleAntigravity:
-		return lower == "antigravity" ||
+		return protocol == "antigravity" || protocol == "google-antigravity"
-			lower == "google-antigravity" ||
-			strings.HasPrefix(lower, "antigravity/") ||
-			strings.HasPrefix(lower, "google-antigravity/")
 	default:
 		return false
 	}
@@ -781,19 +778,22 @@ func defaultModelConfigForProvider(provider, authMethod string) *config.ModelCon
 	case oauthProviderOpenAI:
 		return &config.ModelConfig{
 			ModelName:  "gpt-5.4",
-			Model:      "openai/gpt-5.4",
+			Provider:   "openai",
+			Model:      "gpt-5.4",
 			AuthMethod: authMethod,
 		}
 	case oauthProviderAnthropic:
 		return &config.ModelConfig{
 			ModelName:  "claude-sonnet-4.6",
-			Model:      "anthropic/claude-sonnet-4.6",
+			Provider:   "anthropic",
+			Model:      "claude-sonnet-4.6",
 			AuthMethod: authMethod,
 		}
 	case oauthProviderGoogleAntigravity:
 		return &config.ModelConfig{
 			ModelName:  "gemini-flash",
-			Model:      "antigravity/gemini-3-flash",
+			Provider:   "antigravity",
+			Model:      "gemini-3-flash",
 			AuthMethod: authMethod,
 		}
 	default:

web/backend/api/oauth_test.go

@@ -214,6 +214,54 @@ func TestOAuthLogoutClearsCredentialAndConfig(t *testing.T) {
 	}
 }
 
+func TestOAuthLogoutClearsAuthMethodForExplicitProviderField(t *testing.T) {
+	configPath, cleanup := setupOAuthTestEnv(t)
+	defer cleanup()
+	resetOAuthHooks(t)
+
+	cfg, err := config.LoadConfig(configPath)
+	if err != nil {
+		t.Fatalf("LoadConfig error: %v", err)
+	}
+	cfg.ModelList = append(cfg.ModelList, &config.ModelConfig{
+		ModelName:  "gpt-5.4",
+		Provider:   "openai",
+		Model:      "gpt-5.4",
+		AuthMethod: "oauth",
+	})
+	if err = config.SaveConfig(configPath, cfg); err != nil {
+		t.Fatalf("SaveConfig error: %v", err)
+	}
+	if err = auth.SetCredential(oauthProviderOpenAI, &auth.AuthCredential{
+		AccessToken: "token-before-logout",
+		Provider:    oauthProviderOpenAI,
+		AuthMethod:  "oauth",
+	}); err != nil {
+		t.Fatalf("SetCredential error: %v", err)
+	}
+
+	h := NewHandler(configPath)
+	mux := http.NewServeMux()
+	h.RegisterRoutes(mux)
+
+	rec := httptest.NewRecorder()
+	req := httptest.NewRequest(http.MethodPost, "/api/oauth/logout", bytes.NewBufferString(`{"provider":"openai"}`))
+	req.Header.Set("Content-Type", "application/json")
+	mux.ServeHTTP(rec, req)
+
+	if rec.Code != http.StatusOK {
+		t.Fatalf("status = %d, want %d, body=%s", rec.Code, http.StatusOK, rec.Body.String())
+	}
+
+	updated, err := config.LoadConfig(configPath)
+	if err != nil {
+		t.Fatalf("LoadConfig error: %v", err)
+	}
+	if got := updated.ModelList[len(updated.ModelList)-1].AuthMethod; got != "" {
+		t.Fatalf("auth_method = %q, want empty", got)
+	}
+}
+
 func setupOAuthTestEnv(t *testing.T) (string, func()) {
 	t.Helper()
 

web/frontend/src/api/models.ts

@@ -6,6 +6,7 @@ import { refreshGatewayState } from "@/store/gateway"
 export interface ModelInfo {
   index: number
   model_name: string
+  provider?: string
   model: string
   api_base?: string
   api_key: string

web/frontend/src/components/models/add-model-sheet.tsx

@@ -24,6 +24,7 @@ import { Textarea } from "@/components/ui/textarea"
 
 interface AddForm {
   modelName: string
+  provider: string
   model: string
   apiBase: string
   apiKey: string
@@ -41,6 +42,7 @@ interface AddForm {
 
 const EMPTY_ADD_FORM: AddForm = {
   modelName: "",
+  provider: "",
   model: "",
   apiBase: "",
   apiKey: "",
@@ -119,9 +121,11 @@ export function AddModelSheet({
     setServerError("")
     try {
       const modelName = form.modelName.trim()
+      const provider = form.provider.trim()
       const modelId = form.model.trim()
       await addModel({
         model_name: modelName,
+        provider: provider || undefined,
         model: modelId,
         api_base: form.apiBase.trim() || undefined,
         api_key: form.apiKey.trim() || undefined,
@@ -186,6 +190,17 @@ export function AddModelSheet({
               )}
             </Field>
 
+            <Field
+              label={t("models.field.provider")}
+              hint={t("models.field.providerHint")}
+            >
+              <Input
+                value={form.provider}
+                onChange={setField("provider")}
+                placeholder={t("models.field.providerPlaceholder")}
+              />
+            </Field>
+
             <Field
               label={t("models.add.modelId")}
               hint={t("models.add.modelIdHint")}

web/frontend/src/components/models/edit-model-sheet.tsx

@@ -23,6 +23,8 @@ import {
 import { Textarea } from "@/components/ui/textarea"
 
 interface EditForm {
+  provider: string
+  modelId: string
   apiKey: string
   apiBase: string
   proxy: string
@@ -52,6 +54,8 @@ export function EditModelSheet({
 }: EditModelSheetProps) {
   const { t } = useTranslation()
   const [form, setForm] = useState<EditForm>({
+    provider: "",
+    modelId: "",
     apiKey: "",
     apiBase: "",
     proxy: "",
@@ -72,6 +76,8 @@ export function EditModelSheet({
   useEffect(() => {
     if (model) {
       setForm({
+        provider: model.provider ?? "",
+        modelId: model.model,
         apiKey: "",
         apiBase: model.api_base ?? "",
         proxy: model.proxy ?? "",
@@ -103,12 +109,17 @@ export function EditModelSheet({
 
   const handleSave = async () => {
     if (!model) return
+    if (!form.modelId.trim()) {
+      setError(t("models.add.errorRequired"))
+      return
+    }
     setSaving(true)
     setError("")
     try {
       await updateModel(model.index, {
         model_name: model.model_name,
-        model: model.model,
+        provider: form.provider.trim() || undefined,
+        model: form.modelId.trim(),
         api_base: form.apiBase || undefined,
         api_key: form.apiKey || undefined,
         proxy: form.proxy || undefined,
@@ -166,6 +177,29 @@ export function EditModelSheet({
 
         <div className="min-h-0 flex-1 overflow-y-auto">
           <div className="space-y-5 px-6 py-5">
+            <Field
+              label={t("models.field.provider")}
+              hint={t("models.field.providerHint")}
+            >
+              <Input
+                value={form.provider}
+                onChange={setField("provider")}
+                placeholder={t("models.field.providerPlaceholder")}
+              />
+            </Field>
+
+            <Field
+              label={t("models.add.modelId")}
+              hint={t("models.add.modelIdHint")}
+            >
+              <Input
+                value={form.modelId}
+                onChange={setField("modelId")}
+                placeholder={t("models.add.modelIdPlaceholder")}
+                className="font-mono text-sm"
+              />
+            </Field>
+
             {!isOAuth && (
               <Field
                 label={t("models.field.apiKey")}

web/frontend/src/components/models/models-page.tsx

@@ -20,19 +20,28 @@ const PROVIDER_PRIORITY: Record<string, number> = {
   zhipu: 4,
   deepseek: 5,
   openrouter: 6,
-  qwen: 7,
+  "qwen-portal": 7,
-  moonshot: 8,
+  "qwen-intl": 8,
-  groq: 9,
+  moonshot: 9,
-  "github-copilot": 10,
+  groq: 10,
-  antigravity: 11,
+  "github-copilot": 11,
-  nvidia: 12,
+  antigravity: 12,
-  cerebras: 13,
+  nvidia: 13,
-  shengsuanyun: 14,
+  cerebras: 14,
-  ollama: 15,
+  shengsuanyun: 15,
-  vllm: 16,
+  venice: 16,
-  mistral: 17,
+  vivgrid: 17,
-  avian: 18,
+  minimax: 18,
-  mimo: 19,
+  longcat: 19,
+  modelscope: 20,
+  mistral: 21,
+  avian: 22,
+  azure: 23,
+  ollama: 24,
+  vllm: 25,
+  lmstudio: 26,
+  zai: 27,
+  mimo: 28,
 }
 
 interface ProviderGroup {
@@ -95,10 +104,10 @@ export function ModelsPage() {
 
   const grouped: Record<string, { label: string; models: ModelInfo[] }> = {}
   for (const model of models) {
-    const providerKey = getProviderKey(model.model)
+    const providerKey = getProviderKey(model.provider)
     if (!grouped[providerKey]) {
       grouped[providerKey] = {
-        label: getProviderLabel(model.model),
+        label: getProviderLabel(model.provider),
         models: [],
       }
     }

web/frontend/src/components/models/provider-icon.tsx

@@ -3,9 +3,11 @@ import { useMemo, useState } from "react"
 const PROVIDER_ICON_SLUGS: Record<string, string> = {
   openai: "openai",
   anthropic: "anthropic",
+  azure: "microsoftazure",
   gemini: "googlegemini",
   deepseek: "deepseek",
-  qwen: "alibabacloud",
+  "qwen-portal": "alibabacloud",
+  "qwen-intl": "alibabacloud",
   groq: "groq",
   openrouter: "openrouter",
   nvidia: "nvidia",
@@ -20,9 +22,11 @@ const PROVIDER_ICON_SLUGS: Record<string, string> = {
 const PROVIDER_DOMAINS: Record<string, string> = {
   openai: "openai.com",
   anthropic: "anthropic.com",
+  azure: "azure.com",
   gemini: "gemini.google.com",
   deepseek: "deepseek.com",
-  qwen: "qwenlm.ai",
+  "qwen-portal": "qwenlm.ai",
+  "qwen-intl": "alibabacloud.com",
   moonshot: "moonshot.ai",
   groq: "groq.com",
   openrouter: "openrouter.ai",
@@ -33,11 +37,18 @@ const PROVIDER_DOMAINS: Record<string, string> = {
   antigravity: "antigravity.google",
   "github-copilot": "github.com",
   ollama: "ollama.com",
+  lmstudio: "lmstudio.ai",
   mistral: "mistral.ai",
   avian: "avian.io",
   vllm: "vllm.ai",
   zhipu: "zhipuai.cn",
+  zai: "z.ai",
   mimo: "xiaomi.com",
+  venice: "venice.ai",
+  vivgrid: "vivgrid.com",
+  minimax: "minimaxi.com",
+  longcat: "longcat.chat",
+  modelscope: "modelscope.cn",
 }
 
 interface ProviderIconProps {

web/frontend/src/components/models/provider-label.ts

@@ -1,9 +1,11 @@
 const PROVIDER_LABELS: Record<string, string> = {
   openai: "OpenAI",
   anthropic: "Anthropic",
+  azure: "Azure OpenAI",
   gemini: "Google Gemini",
   deepseek: "DeepSeek",
-  qwen: "Qwen (阿里云)",
+  "qwen-portal": "Qwen (阿里云)",
+  "qwen-intl": "Qwen International",
   moonshot: "Moonshot (月之暗面)",
   groq: "Groq",
   openrouter: "OpenRouter",
@@ -14,21 +16,37 @@ const PROVIDER_LABELS: Record<string, string> = {
   antigravity: "Google Code Assist",
   "github-copilot": "GitHub Copilot",
   ollama: "Ollama (local)",
+  lmstudio: "LM Studio (local)",
   mistral: "Mistral AI",
   avian: "Avian",
   vllm: "VLLM (local)",
   zhipu: "Zhipu AI (智谱)",
+  zai: "Z.ai",
   mimo: "Xiaomi MiMo",
+  venice: "Venice AI",
+  vivgrid: "Vivgrid",
+  minimax: "MiniMax",
+  longcat: "LongCat",
+  modelscope: "ModelScope (魔搭社区)",
 }
 
-export function getProviderKey(model: string): string {
+const PROVIDER_ALIASES: Record<string, string> = {
-  return model.split("/")[0]
+  qwen: "qwen-portal",
+  "qwen-international": "qwen-intl",
+  "dashscope-intl": "qwen-intl",
+  "z.ai": "zai",
+  "z-ai": "zai",
+  google: "gemini",
+  "google-antigravity": "antigravity",
 }
 
-export function getProviderLabel(model: string): string {
+export function getProviderKey(provider?: string): string {
-  const prefix = getProviderKey(model)
+  const normalized = provider?.trim().toLowerCase()
-  const labels: Record<string, string> = {
+  if (!normalized) return "openai"
-    ...PROVIDER_LABELS,
+  return PROVIDER_ALIASES[normalized] ?? normalized
-  }
+}
-  return labels[prefix] ?? prefix
+
+export function getProviderLabel(provider?: string): string {
+  const prefix = getProviderKey(provider)
+  return PROVIDER_LABELS[prefix] ?? prefix
 }

web/frontend/src/i18n/locales/en.json

@@ -239,8 +239,8 @@
       "modelNamePlaceholder": "e.g. my-gpt4",
       "modelNameHint": "A short name used to identify this model in conversations.",
       "modelId": "Model Identifier",
-      "modelIdPlaceholder": "e.g. openai/gpt-4o",
+      "modelIdPlaceholder": "e.g. gpt-4o or openai/gpt-4o",
-      "modelIdHint": "Format: protocol/model-id. Supported: openai, anthropic, gemini, groq, …",
+      "modelIdHint": "If Provider is not specified, values such as openai/gpt-4o are interpreted using the provider/model format. If Provider is specified, this field is treated as the canonical model ID and is not parsed for a provider prefix.",
       "errorRequired": "This field is required.",
       "errorDuplicateModelName": "Model alias already exists. Please use a different name.",
       "saveError": "Failed to add model",
@@ -255,6 +255,9 @@
       "toggle": "Advanced options"
     },
     "field": {
+      "provider": "Provider",
+      "providerPlaceholder": "e.g. openai",
+      "providerHint": "Optional. If specified, this value is used as the effective provider, and Model Identifier is interpreted as the canonical model ID.",
       "apiBase": "API Base URL",
       "apiKey": "API Key",
       "apiKeyPlaceholder": "Enter your API key",

web/frontend/src/i18n/locales/zh.json

@@ -239,8 +239,8 @@
       "modelNamePlaceholder": "例如 my-gpt4",
       "modelNameHint": "用于在对话中识别此模型的简短名称。",
       "modelId": "模型标识符",
-      "modelIdPlaceholder": "例如 openai/gpt-4o",
+      "modelIdPlaceholder": "例如 gpt-4o 或 openai/gpt-4o",
-      "modelIdHint": "格式：协议/模型ID。支持：openai、anthropic、gemini、groq 等。",
+      "modelIdHint": "未指定 Provider 时，诸如 openai/gpt-4o 的值将按 provider/model 格式解析。已指定 Provider 时，此字段将作为规范模型 ID 使用，不再解析其中的 provider 前缀。",
       "errorRequired": "此字段为必填项。",
       "errorDuplicateModelName": "模型别名已存在，请使用其他名称。",
       "saveError": "添加模型失败",
@@ -255,6 +255,9 @@
       "toggle": "高级选项"
     },
     "field": {
+      "provider": "Provider",
+      "providerPlaceholder": "例如 openai",
+      "providerHint": "可选。指定后，将以该值作为最终 provider，并将“模型标识符”字段解释为规范模型 ID。",
       "apiBase": "API Base URL",
       "apiKey": "API Key",
       "apiKeyPlaceholder": "请输入 API Key",