update relay-gemini-native.go

relay/channel/gemini/relay-gemini-native.go

@@ -51,7 +51,7 @@ func GeminiTextGenerationHandler(c *gin.Context, resp *http.Response, info *rela
 	// 计算使用量（基于 UsageMetadata）
 	usage := dto.Usage{
 		PromptTokens:     geminiResponse.UsageMetadata.PromptTokenCount,
-		CompletionTokens: geminiResponse.UsageMetadata.CandidatesTokenCount,
+		CompletionTokens: geminiResponse.UsageMetadata.CandidatesTokenCount + geminiResponse.UsageMetadata.ThoughtsTokenCount,
 		TotalTokens:      geminiResponse.UsageMetadata.TotalTokenCount,
 	}
 
@@ -64,9 +64,6 @@ func GeminiTextGenerationHandler(c *gin.Context, resp *http.Response, info *rela
 			usage.PromptTokensDetails.TextTokens = detail.TokenCount
 		}
 	}
-	
-	// 计算最终使用量
-	usage.CompletionTokens = usage.TotalTokens - usage.PromptTokens
 
 	// 直接返回 Gemini 原生格式的 JSON 响应
 	jsonResponse, err := json.Marshal(geminiResponse)
@@ -111,7 +108,7 @@ func GeminiTextGenerationStreamHandler(c *gin.Context, resp *http.Response, info
 		// 更新使用量统计
 		if geminiResponse.UsageMetadata.TotalTokenCount != 0 {
 			usage.PromptTokens = geminiResponse.UsageMetadata.PromptTokenCount
-			usage.CompletionTokens = geminiResponse.UsageMetadata.CandidatesTokenCount
+			usage.CompletionTokens = geminiResponse.UsageMetadata.CandidatesTokenCount + geminiResponse.UsageMetadata.ThoughtsTokenCount
 			usage.TotalTokens = geminiResponse.UsageMetadata.TotalTokenCount
 			usage.CompletionTokenDetails.ReasoningTokens = geminiResponse.UsageMetadata.ThoughtsTokenCount
 			for _, detail := range geminiResponse.UsageMetadata.PromptTokensDetails {
@@ -139,7 +136,7 @@ func GeminiTextGenerationStreamHandler(c *gin.Context, resp *http.Response, info
 	}
 
 	// 计算最终使用量
-	usage.CompletionTokens = usage.TotalTokens - usage.PromptTokens
+	// usage.CompletionTokens = usage.TotalTokens - usage.PromptTokens
 
 	// 移除流式响应结尾的[Done]，因为Gemini API没有发送Done的行为
 	//helper.Done(c)