Drop ZHANLU_MODELS and ZHANLU_DEFAULT_MODEL config

Model list is always fetched live from the gateway model-info endpoint, so the
static fallback list was dead config with stale IDs (GLM-4.7 vs zhanlu/glm-4.7).
Default to zhanlu/auto when the request omits model.

internal/server/server.go

@@ -525,7 +525,7 @@ func (s *Server) exchangeSSOCode(w http.ResponseWriter, r *http.Request) {
 }
 
 func (s *Server) models(w http.ResponseWriter, r *http.Request) {
-	modelIDs := s.cfg.Models
+	modelIDs := []string{}
 	if creds, err := s.currentCredentials(); err == nil && creds.HasAPIKey() {
 		if client, cerr := s.zhanluClient(); cerr == nil {
 			if fetched, merr := client.Models(r.Context(), creds.APIKey); merr == nil && len(fetched) > 0 {
@@ -552,7 +552,7 @@ func (s *Server) chatCompletions(w http.ResponseWriter, r *http.Request) {
 		return
 	}
 	if req.Model == "" {
-		req.Model = s.cfg.DefaultModel
+		req.Model = "zhanlu/auto"
 	}
 	clientWantsStream := req.Stream
 	req.Stream = true

internal/server/server_test.go

@@ -70,8 +70,6 @@ func setupTestServer(t *testing.T) (*httptest.Server, *httptest.Server, string)
 		PublicKeyPEM:       defaultTestPublicKey,
 		PhonePublicKeyPEM:  defaultTestPublicKey,
 		SM2PrivateKey:      testSM2Key,
-		Models:             []string{"GLM-4.7", "MiniMaxAI/MiniMax-M2.5"},
-		DefaultModel:       "GLM-4.7",
 		PluginVersion:      "1.4.2",
 	}
 	h := New(cfg)