feat: add image generation for OpenAI DALL-E and Gemini Imagen

New `/v1/images/generations` endpoint proxies DALL-E 2/3 (OpenAI)
and Imagen 3 (Gemini). Same auth/logging as chat completions.

- Add ImageGenerationRequest/Response models
- Extend Provider interface with ImageGeneration()
- OpenAI: forward to /v1/images/generations
- Gemini: call /v1beta/models/{model}:predict, map OpenAI params
- Circuit breaker wraps image gen like chat completions
- Model routing: dall-e* -> openai, imagen*/gemini* -> gemini
- Unsupported providers (deepseek/moonshot/grok/ollama) return error
- Fix pre-existing CachedContentTokenCount bug in StreamGemini

internal/server/server.go

@@ -186,6 +186,7 @@ func (s *Server) setupRoutes() {
 	v1.Use(middleware.AuthMiddleware(s.database, true))
 	{
 		v1.POST("/chat/completions", s.handleChatCompletions)
+		v1.POST("/images/generations", s.handleImageGenerations)
 		v1.GET("/models", s.handleListModels)
 		v1.GET("/responses", s.handleListResponses)
 	}
@@ -501,6 +502,71 @@ func (s *Server) handleChatCompletions(c *gin.Context) {
 	c.JSON(http.StatusOK, resp)
 }
 
+func (s *Server) handleImageGenerations(c *gin.Context) {
+	startTime := time.Now()
+	var req models.ImageGenerationRequest
+	if err := c.ShouldBindJSON(&req); err != nil {
+		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
+		return
+	}
+
+	// Determine provider based on model name
+	providerName := "openai"
+	modelLower := strings.ToLower(req.Model)
+	switch {
+	case strings.Contains(modelLower, "imagen"), strings.Contains(modelLower, "gemini"):
+		providerName = "gemini"
+	case strings.Contains(modelLower, "dall"), strings.HasPrefix(modelLower, "openai/"):
+		providerName = "openai"
+	}
+
+	// Default model for each provider if not specified
+	if req.Model == "" {
+		if providerName == "openai" {
+			req.Model = "dall-e-3"
+		} else {
+			req.Model = "imagen-3.0-generate-001"
+		}
+	}
+
+	// Strip common prefixes
+	prefixes := []string{"openai/", "gemini/", "google/"}
+	for _, p := range prefixes {
+		if strings.HasPrefix(req.Model, p) {
+			req.Model = strings.TrimPrefix(req.Model, p)
+			break
+		}
+	}
+
+	provider, ok := s.providers[providerName]
+	if !ok {
+		c.JSON(http.StatusInternalServerError, gin.H{"error": fmt.Sprintf("Provider %s not enabled or supported", providerName)})
+		return
+	}
+
+	clientID := "default"
+	if auth, ok := c.Get("auth"); ok {
+		if authInfo, ok := auth.(models.AuthInfo); ok {
+			clientID = authInfo.ClientID
+		}
+	}
+
+	resp, err := provider.ImageGeneration(c.Request.Context(), &req)
+	if err != nil {
+		s.logRequest(startTime, clientID, providerName, req.Model, nil, err, false)
+		c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
+		return
+	}
+
+	s.logRequest(startTime, clientID, providerName, req.Model, &models.Usage{
+		PromptTokens:     1,
+		CompletionTokens: uint32(len(resp.Data)),
+		TotalTokens:      1 + uint32(len(resp.Data)),
+	}, nil, false)
+
+	c.JSON(http.StatusOK, resp)
+}
+
 func (s *Server) logRequest(start time.Time, clientID, provider, model string, usage *models.Usage, err error, hasImages bool) {
 	entry := RequestLog{
 		Timestamp:  start,