diff --git a/internal/entity/models/gitee.go b/internal/entity/models/gitee.go
index 8ba04c792c..976cdf4427 100644
--- a/internal/entity/models/gitee.go
+++ b/internal/entity/models/gitee.go
@@ -53,34 +53,6 @@ func (g *GiteeModel) Name() string {
 	return "GiteeAI"
 }
 
-// GiteeChatResponse mirrors GiteeAI's OpenAI-compatible chat response.
-type GiteeChatResponse struct {
-	ID      string `json:"id"`
-	Object  string `json:"object"`
-	Created int64  `json:"created"`
-	Model   string `json:"model"`
-	Choices []struct {
-		FinishReason string `json:"finish_reason"`
-		Index        int    `json:"index"`
-		Logprobs     any    `json:"logprobs"`
-		Message      struct {
-			Content          string           `json:"content"`
-			ReasoningContent string           `json:"reasoning_content"`
-			Role             string           `json:"role"`
-			ToolCalls        []map[string]any `json:"tool_calls"`
-		} `json:"message"`
-	} `json:"choices"`
-	SystemFingerprint string `json:"system_fingerprint"`
-	Usage             struct {
-		CompletionTokens    int `json:"completion_tokens"`
-		PromptTokens        int `json:"prompt_tokens"`
-		TotalTokens         int `json:"total_tokens"`
-		PromptTokensDetails struct {
-			CachedTokens int `json:"cached_tokens"`
-		} `json:"prompt_tokens_details"`
-	} `json:"usage"`
-}
-
 // ChatWithMessages sends multiple messages with roles and returns response
 func (g *GiteeModel) ChatWithMessages(ctx context.Context, modelName string, messages []Message, apiConfig *APIConfig, chatModelConfig *ChatConfig, modelUsage *common.ModelUsage) (*ChatResponse, error) {
 	if err := g.baseModel.APIConfigCheck(apiConfig); err != nil {
@@ -112,81 +84,12 @@ func (g *GiteeModel) ChatWithMessages(ctx context.Context, modelName string, mes
 		}
 	}
 
-	jsonData, err := json.Marshal(reqBody)
+	body, err := g.baseModel.doRequest(ctx, url, apiConfig, reqBody, nonStreamCallTimeout)
 	if err != nil {
-		return nil, fmt.Errorf("failed to marshal request: %w", err)
+		return nil, err
 	}
 
-	common.Info(fmt.Sprintf("GiteeAPI request body: %s", string(jsonData)))
-
-	ctx, cancel := context.WithTimeout(ctx, nonStreamCallTimeout)
-	defer cancel()
-
-	req, err := http.NewRequestWithContext(ctx, "POST", url, bytes.NewBuffer(jsonData))
-	if err != nil {
-		return nil, fmt.Errorf("failed to create request: %w", err)
-	}
-
-	req.Header.Set("Content-Type", "application/json")
-	req.Header.Set("Authorization", fmt.Sprintf("Bearer %s", *apiConfig.ApiKey))
-
-	resp, err := g.baseModel.httpClient.Do(req)
-	if err != nil {
-		return nil, fmt.Errorf("failed to send request: %w", err)
-	}
-	defer resp.Body.Close()
-
-	body, err := io.ReadAll(resp.Body)
-	if err != nil {
-		return nil, fmt.Errorf("failed to read response: %w", err)
-	}
-
-	if resp.StatusCode != http.StatusOK {
-		return nil, fmt.Errorf("API request failed with status %d: %s", resp.StatusCode, string(body))
-	}
-
-	return parseChatCompletionResponse(body, chatModelConfig, modelUsage, func(body []byte, chatConfig *ChatConfig) (chatResponseParts, error) {
-		var result GiteeChatResponse
-		if err := json.Unmarshal(body, &result); err != nil {
-			return chatResponseParts{}, fmt.Errorf("failed to parse response: %w", err)
-		}
-		if len(result.Choices) == 0 {
-			return chatResponseParts{}, fmt.Errorf("no choices in response")
-		}
-
-		choice := result.Choices[0]
-		content := choice.Message.Content
-		if content == "" && len(choice.Message.ToolCalls) == 0 {
-			return chatResponseParts{}, fmt.Errorf("response contains neither content nor tool calls")
-		}
-
-		reasonContent := ""
-		if chatConfig != nil && chatConfig.Thinking != nil && *chatConfig.Thinking {
-			reasonContent = choice.Message.ReasoningContent
-			if reasonContent == "" {
-				reasoning, answer := GetThinkingAndAnswer(chatConfig.ModelClass, &content)
-				if reasoning != nil {
-					reasonContent = *reasoning
-					content = *answer
-				}
-			}
-			if strings.HasPrefix(reasonContent, "\n") {
-				reasonContent = reasonContent[1:]
-			}
-		}
-
-		return chatResponseParts{
-			RequestID:     result.ID,
-			Content:       &content,
-			ReasonContent: &reasonContent,
-			ToolCalls:     choice.Message.ToolCalls,
-			Usage: &TokenUsage{
-				PromptTokens:     result.Usage.PromptTokens,
-				CompletionTokens: result.Usage.CompletionTokens,
-				TotalTokens:      result.Usage.TotalTokens,
-			},
-		}, nil
-	})
+	return HandleNonStreamingResponse(body, modelUsage, chatModelConfig, OpenAIParserConfig)
 }
 
 // ChatStreamlyWithSender sends messages and streams response via sender function (best performance, no channel)
@@ -226,128 +129,9 @@ func (g *GiteeModel) ChatStreamlyWithSender(ctx context.Context, modelName strin
 
 	}
 
-	jsonData, err := json.Marshal(reqBody)
-	if err != nil {
-		return fmt.Errorf("failed to marshal request: %w", err)
-	}
-
-	ctx, cancel := context.WithTimeout(ctx, streamCallTimeout)
-	defer cancel()
-
-	req, err := http.NewRequestWithContext(ctx, "POST", url, bytes.NewBuffer(jsonData))
-	if err != nil {
-		return fmt.Errorf("failed to create request: %w", err)
-	}
-
-	req.Header.Set("Content-Type", "application/json")
-	req.Header.Set("Authorization", fmt.Sprintf("Bearer %s", *apiConfig.ApiKey))
-
-	resp, err := g.baseModel.httpClient.Do(req)
-	if err != nil {
-		return fmt.Errorf("failed to send request: %w", err)
-	}
-	defer resp.Body.Close()
-
-	if resp.StatusCode != http.StatusOK {
-		body, _ := io.ReadAll(resp.Body)
-		return fmt.Errorf("API request failed with status %d: %s", resp.StatusCode, string(body))
-	}
-
-	reserveText := ""
-	thinkingPhase := false
-	answerPhase := false
-	sawTerminal := false
-
-	accumulatedToolCalls := make(map[int]map[string]any)
-	done, err := ParseSSEStream[map[string]any](resp.Body, func(event map[string]any) error {
-		common.Info(fmt.Sprintf("%v", event))
-
-		tokenUsage, found, usageErr := decodeOpenAICompatibleStreamUsage(event)
-		if usageErr != nil {
-			return usageErr
-		}
-		if found {
-			applyStreamUsage(chatModelConfig, modelUsage, tokenUsage)
-		}
-		choices, ok := event["choices"].([]any)
-		if !ok || len(choices) == 0 {
-			return nil
-		}
-		choice, ok := choices[0].(map[string]any)
-		if !ok {
-			return nil
-		}
-		if finishReason, ok := choice["finish_reason"].(string); ok && finishReason != "" {
-			sawTerminal = true
-		}
-		delta, ok := choice["delta"].(map[string]any)
-		if !ok {
-			return nil
-		}
-
-		accumulateToolCallDeltas(delta, accumulatedToolCalls)
-		if reasoning, ok := delta["reasoning_content"].(string); ok && reasoning != "" {
-			if err := sender(nil, &reasoning); err != nil {
-				return err
-			}
-		}
-
-		content, ok := delta["content"].(string)
-		if ok && content != "" {
-			common.Info(content)
-			if content == "<think>" {
-				thinkingPhase = true
-				return nil
-
-			} else if content == "</think>" {
-				thinkingPhase = false
-				answerPhase = true
-				return nil
-			}
-
-			if thinkingPhase {
-				if err = sender(nil, &content); err != nil {
-					return err
-				}
-				reserveText = ""
-			} else if answerPhase {
-				if err = sender(&content, nil); err != nil {
-					return err
-				}
-				reserveText = ""
-			} else {
-				content = strings.Trim(content, "\n")
-				content = strings.Trim(content, " ")
-				if content != "" {
-					reserveText += content
-				}
-			}
-		}
-
-		return nil
+	return g.baseModel.doStreamRequest(ctx, url, apiConfig, reqBody, streamCallTimeout, func(body io.ReadCloser) error {
+		return HandleStreamingResponse(body, modelUsage, chatModelConfig, OpenAIParserConfig, sender)
 	})
-	if err != nil {
-		return fmt.Errorf("failed to scan response body: %w", err)
-	}
-	if !done && !sawTerminal {
-		return fmt.Errorf("gitee: stream ended before [DONE] or finish_reason")
-	}
-
-	if reserveText != "" {
-		if err = sender(&reserveText, nil); err != nil {
-			return err
-		}
-	}
-
-	setSortedToolCallsResult(chatModelConfig, accumulatedToolCalls)
-
-	// Send [DONE] marker for OpenAI compatibility
-	endOfStream := "[DONE]"
-	if err = sender(&endOfStream, nil); err != nil {
-		return err
-	}
-
-	return nil
 }
 
 // GiteeEmbeddingResponse mirrors GiteeAI's embeddings response.
diff --git a/internal/entity/models/gitee_test.go b/internal/entity/models/gitee_test.go
index 280fefcae4..957e9fd1ff 100644
--- a/internal/entity/models/gitee_test.go
+++ b/internal/entity/models/gitee_test.go
@@ -57,7 +57,8 @@ func TestGiteeStreamAcceptsTerminalWithoutDelta(t *testing.T) {
 			t.Errorf("method=%s, want POST", r.Method)
 		}
 		w.Header().Set("Content-Type", "text/event-stream")
-		_, _ = io.WriteString(w, `data: {"choices":[{"finish_reason":"stop"}]}`+"\n\n")
+		_, _ = io.WriteString(w, `data: {"choices":[{"finish_reason":"stop"}]}`+"\n\n"+
+			`data: [DONE]`+"\n\n")
 	}))
 	defer srv.Close()
 
diff --git a/internal/entity/models/gpustack.go b/internal/entity/models/gpustack.go
index fb87f4afde..cb4bbf598c 100644
--- a/internal/entity/models/gpustack.go
+++ b/internal/entity/models/gpustack.go
@@ -71,70 +71,12 @@ func (g *GPUStackModel) ChatWithMessages(ctx context.Context, modelName string,
 	url := fmt.Sprintf("%s/%s", baseURL, g.baseModel.URLSuffix.Chat)
 	reqBody := buildRequestBody(chatModelConfig, modelName, messages, false)
 
-	jsonData, err := json.Marshal(reqBody)
+	body, err := g.baseModel.doRequest(ctx, url, apiConfig, reqBody, nonStreamCallTimeout)
 	if err != nil {
-		return nil, fmt.Errorf("failed to marshal request: %w", err)
+		return nil, err
 	}
 
-	ctx, cancel := context.WithTimeout(ctx, nonStreamCallTimeout)
-	defer cancel()
-
-	req, err := http.NewRequestWithContext(ctx, http.MethodPost, url, bytes.NewBuffer(jsonData))
-	if err != nil {
-		return nil, fmt.Errorf("failed to create request: %w", err)
-	}
-	req.Header.Set("Content-Type", "application/json")
-	if auth := BearerAuth(apiConfig); auth != "" {
-		req.Header.Set("Authorization", auth)
-	}
-
-	resp, err := g.baseModel.httpClient.Do(req)
-	if err != nil {
-		return nil, fmt.Errorf("failed to send request: %w", err)
-	}
-	defer resp.Body.Close()
-
-	body, err := io.ReadAll(resp.Body)
-	if err != nil {
-		return nil, fmt.Errorf("failed to read response: %w", err)
-	}
-	if resp.StatusCode != http.StatusOK {
-		return nil, fmt.Errorf("API request failed with status %d: %s", resp.StatusCode, string(body))
-	}
-
-	var result map[string]interface{}
-	if err = json.Unmarshal(body, &result); err != nil {
-		return nil, fmt.Errorf("failed to parse response: %w", err)
-	}
-
-	choices, ok := result["choices"].([]interface{})
-	if !ok || len(choices) == 0 {
-		return nil, fmt.Errorf("no choices in response")
-	}
-	firstChoice, ok := choices[0].(map[string]interface{})
-	if !ok {
-		return nil, fmt.Errorf("invalid choice format")
-	}
-	messageMap, ok := firstChoice["message"].(map[string]interface{})
-	if !ok {
-		return nil, fmt.Errorf("invalid message format")
-	}
-	content, ok := messageMap["content"].(string)
-	toolCalls := extractToolCalls(messageMap)
-	if !ok && len(toolCalls) == 0 {
-		return nil, fmt.Errorf("invalid content format")
-	}
-
-	reasonContent := ""
-	if r, ok := messageMap["reasoning_content"].(string); ok {
-		reasonContent = r
-	}
-
-	return &ChatResponse{
-		Answer:        &content,
-		ReasonContent: &reasonContent,
-		ToolCalls:     toolCalls,
-	}, nil
+	return HandleNonStreamingResponse(body, modelUsage, chatModelConfig, OpenAIParserConfig)
 }
 
 // ChatStreamlyWithSender streams the response via the sender.
@@ -163,12 +105,16 @@ func (g *GPUStackModel) ChatStreamlyWithSender(ctx context.Context, modelName st
 	baseURL = strings.TrimSuffix(baseURL, "/")
 	url := fmt.Sprintf("%s/%s", baseURL, g.baseModel.URLSuffix.Chat)
 	reqBody := buildRequestBody(chatModelConfig, modelName, messages, true)
+	reqBody["stream_options"] = map[string]any{"include_usage": true}
 
 	jsonData, err := json.Marshal(reqBody)
 	if err != nil {
 		return fmt.Errorf("failed to marshal request: %w", err)
 	}
 
+	ctx, cancel := context.WithTimeout(ctx, streamCallTimeout)
+	defer cancel()
+
 	req, err := http.NewRequestWithContext(ctx, http.MethodPost, url, bytes.NewBuffer(jsonData))
 	if err != nil {
 		return fmt.Errorf("failed to create request: %w", err)
@@ -190,60 +136,7 @@ func (g *GPUStackModel) ChatStreamlyWithSender(ctx context.Context, modelName st
 		return fmt.Errorf("API request failed with status %d: %s", resp.StatusCode, string(body))
 	}
 
-	sawTerminal := false
-	accumulatedToolCalls := make(map[int]map[string]any)
-	done, err := ParseSSEStream[map[string]interface{}](resp.Body, func(event map[string]interface{}) error {
-		if apiErr, ok := event["error"]; ok {
-			return fmt.Errorf("gpustack: upstream stream error: %v", apiErr)
-		}
-
-		choices, ok := event["choices"].([]interface{})
-		if !ok || len(choices) == 0 {
-			return nil
-		}
-		firstChoice, ok := choices[0].(map[string]interface{})
-		if !ok {
-			return nil
-		}
-		delta, ok := firstChoice["delta"].(map[string]interface{})
-		if !ok {
-			return nil
-		}
-
-		accumulateToolCallDeltas(delta, accumulatedToolCalls)
-
-		if r, ok := delta["reasoning_content"].(string); ok && r != "" {
-			if err := sender(nil, &r); err != nil {
-				return err
-			}
-		}
-
-		content, ok := delta["content"].(string)
-		if ok && content != "" {
-			if err := sender(&content, nil); err != nil {
-				return err
-			}
-		}
-
-		finishReason, ok := firstChoice["finish_reason"].(string)
-		if ok && finishReason != "" {
-			sawTerminal = true
-		}
-		return nil
-	})
-	if err != nil {
-		return fmt.Errorf("failed to scan response body: %w", err)
-	}
-	setSortedToolCallsResult(chatModelConfig, accumulatedToolCalls)
-	if !done && !sawTerminal {
-		return fmt.Errorf("gpustack: stream ended before [DONE] or finish_reason")
-	}
-
-	endOfStream := "[DONE]"
-	if err := sender(&endOfStream, nil); err != nil {
-		return err
-	}
-	return nil
+	return HandleStreamingResponse(resp.Body, modelUsage, chatModelConfig, OpenAIParserConfig, sender)
 }
 
 type gpustackModelInfo struct {
diff --git a/internal/entity/models/greenpt.go b/internal/entity/models/greenpt.go
index baa57244b2..3711d293c5 100644
--- a/internal/entity/models/greenpt.go
+++ b/internal/entity/models/greenpt.go
@@ -67,6 +67,88 @@ func (m *GreenPTModel) ListModels(ctx context.Context, apiConfig *APIConfig) ([]
 	return models, nil
 }
 
+// ChatWithMessages sends multiple messages with roles and returns response
+func (m *GreenPTModel) ChatWithMessages(ctx context.Context, modelName string, messages []Message, apiConfig *APIConfig, chatModelConfig *ChatConfig, modelUsage *common.ModelUsage) (*ChatResponse, error) {
+	if err := m.baseModel.APIConfigCheck(apiConfig); err != nil {
+		return nil, err
+	}
+
+	if len(messages) == 0 {
+		return nil, fmt.Errorf("messages is empty")
+	}
+
+	resolvedBaseURL, err := m.baseModel.GetBaseURL(apiConfig)
+	if err != nil {
+		return nil, err
+	}
+	url := fmt.Sprintf("%s/%s", resolvedBaseURL, m.baseModel.URLSuffix.Chat)
+
+	// Build request body
+	reqBody := buildRequestBody(chatModelConfig, modelName, messages, false)
+
+	if chatModelConfig != nil && chatModelConfig.Thinking != nil {
+		if *chatModelConfig.Thinking {
+			reqBody["thinking"] = map[string]interface{}{
+				"type": "enabled",
+			}
+		} else {
+			reqBody["thinking"] = map[string]interface{}{
+				"type": "disabled",
+			}
+		}
+	}
+
+	body, err := m.baseModel.doRequest(ctx, url, apiConfig, reqBody, nonStreamCallTimeout)
+	if err != nil {
+		return nil, err
+	}
+
+	return HandleNonStreamingResponse(body, modelUsage, chatModelConfig, OpenAIParserConfig)
+}
+
+// ChatStreamlyWithSender sends messages and streams response via sender function
+func (m *GreenPTModel) ChatStreamlyWithSender(ctx context.Context, modelName string, messages []Message, apiConfig *APIConfig, chatModelConfig *ChatConfig, modelUsage *common.ModelUsage, sender func(*string, *string) error) error {
+	if err := m.baseModel.APIConfigCheck(apiConfig); err != nil {
+		return err
+	}
+
+	if len(messages) == 0 {
+		return fmt.Errorf("messages is empty")
+	}
+	if chatModelConfig != nil {
+		chatModelConfig.ToolCallsResult = nil
+		chatModelConfig.UsageResult = nil
+	}
+
+	resolvedBaseURL, err := m.baseModel.GetBaseURL(apiConfig)
+	if err != nil {
+		return err
+	}
+	url := fmt.Sprintf("%s/%s", resolvedBaseURL, m.baseModel.URLSuffix.Chat)
+
+	// Build request body with streaming enabled
+	reqBody := buildRequestBody(chatModelConfig, modelName, messages, true)
+	reqBody["stream_options"] = map[string]interface{}{
+		"include_usage": true,
+	}
+
+	if chatModelConfig != nil && chatModelConfig.Thinking != nil {
+		if *chatModelConfig.Thinking {
+			reqBody["thinking"] = map[string]interface{}{
+				"type": "enabled",
+			}
+		} else {
+			reqBody["thinking"] = map[string]interface{}{
+				"type": "disabled",
+			}
+		}
+	}
+
+	return m.baseModel.doStreamRequest(ctx, url, apiConfig, reqBody, streamCallTimeout, func(body io.ReadCloser) error {
+		return HandleStreamingResponse(body, modelUsage, chatModelConfig, OpenAIParserConfig, sender)
+	})
+}
+
 // TranscribeAudio calls GreenPT's Deepgram-compatible /v1/listen endpoint.
 func (m *GreenPTModel) TranscribeAudio(ctx context.Context, modelName *string, file *string, apiConfig *APIConfig, asrConfig *ASRConfig, modelUsage *common.ModelUsage) (*ASRResponse, error) {
 	if err := m.baseModel.APIConfigCheck(apiConfig); err != nil {
diff --git a/internal/entity/models/huaweicloud.go b/internal/entity/models/huaweicloud.go
index a032463c45..4368348095 100644
--- a/internal/entity/models/huaweicloud.go
+++ b/internal/entity/models/huaweicloud.go
@@ -50,37 +50,6 @@ func (h *HuaweiCloudModel) Name() string {
 	return "huaweicloud"
 }
 
-// HuaweiCloudChatResponse captures the OpenAI-compatible fields consumed by RAGFlow.
-type HuaweiCloudChatResponse struct {
-	ID      string `json:"id"`
-	Object  string `json:"object"`
-	Created int64  `json:"created"`
-	Model   string `json:"model"`
-	Choices []struct {
-		FinishReason string `json:"finish_reason"`
-		Index        int    `json:"index"`
-		Logprobs     any    `json:"logprobs"`
-		Message      struct {
-			Content          string           `json:"content"`
-			ReasoningContent string           `json:"reasoning_content"`
-			Role             string           `json:"role"`
-			ToolCalls        []map[string]any `json:"tool_calls"`
-		} `json:"message"`
-	} `json:"choices"`
-	SystemFingerprint string `json:"system_fingerprint"`
-	Usage             struct {
-		CompletionTokens        int `json:"completion_tokens"`
-		PromptTokens            int `json:"prompt_tokens"`
-		TotalTokens             int `json:"total_tokens"`
-		CompletionTokensDetails struct {
-			ReasoningTokens int `json:"reasoning_tokens"`
-		} `json:"completion_tokens_details"`
-		PromptTokensDetails struct {
-			CachedTokens int `json:"cached_tokens"`
-		} `json:"prompt_tokens_details"`
-	} `json:"usage"`
-}
-
 func huaweiCloudRegion(api *APIConfig) string {
 	region := "default"
 	if api != nil && api.Region != nil && *api.Region != "" {
@@ -222,73 +191,12 @@ func (h *HuaweiCloudModel) ChatWithMessages(ctx context.Context, modelName strin
 	huaweiCloudApplyChatConfig(reqb, modelName, chatModelConfig)
 	applyChatToolConfig(reqb, chatModelConfig)
 
-	jsonData, err := json.Marshal(reqb)
+	body, err := h.baseModel.doRequest(ctx, url, apiConfig, reqb, nonStreamCallTimeout)
 	if err != nil {
-		return nil, fmt.Errorf("failed to marshal request: %w", err)
+		return nil, err
 	}
 
-	ctx, cancel := context.WithTimeout(ctx, nonStreamCallTimeout)
-	defer cancel()
-
-	req, err := http.NewRequestWithContext(ctx, "POST", url, bytes.NewBuffer(jsonData))
-	if err != nil {
-		return nil, fmt.Errorf("failed to create request: %w", err)
-	}
-	req.Header.Set("Content-Type", "application/json")
-	req.Header.Set("Authorization", huaweiCloudAuthorization(*apiConfig.ApiKey))
-
-	rep, err := h.baseModel.httpClient.Do(req)
-	if err != nil {
-		return nil, fmt.Errorf("failed to send request: %w", err)
-	}
-	defer rep.Body.Close()
-
-	body, err := io.ReadAll(rep.Body)
-	if err != nil {
-		return nil, fmt.Errorf("failed to read response: %w", err)
-	}
-	if rep.StatusCode != http.StatusOK {
-		return nil, fmt.Errorf("Huawei Cloud chat API error: status %d, body: %s", rep.StatusCode, string(body))
-	}
-
-	return parseChatCompletionResponse(body, chatModelConfig, modelUsage, func(body []byte, chatConfig *ChatConfig) (chatResponseParts, error) {
-		var result HuaweiCloudChatResponse
-		if err := json.Unmarshal(body, &result); err != nil {
-			return chatResponseParts{}, fmt.Errorf("failed to parse response: %w", err)
-		}
-		if len(result.Choices) == 0 {
-			return chatResponseParts{}, fmt.Errorf("no choices in response")
-		}
-
-		choice := &result.Choices[0]
-		content := choice.Message.Content
-		if content == "" && len(choice.Message.ToolCalls) == 0 {
-			return chatResponseParts{}, fmt.Errorf("invalid content format")
-		}
-		reasonContent := choice.Message.ReasoningContent
-		if chatConfig != nil && chatConfig.Thinking != nil && *chatConfig.Thinking && reasonContent == "" {
-			reasoning, answer := GetThinkingAndAnswer(chatConfig.ModelClass, &content)
-			if reasoning != nil {
-				reasonContent = *reasoning
-				content = *answer
-			}
-		}
-		if reasonContent != "" && reasonContent[0] == '\n' {
-			reasonContent = reasonContent[1:]
-		}
-
-		return chatResponseParts{
-			RequestID:     result.ID,
-			Content:       &content,
-			ReasonContent: &reasonContent,
-			ToolCalls:     choice.Message.ToolCalls,
-			Usage: &TokenUsage{
-				PromptTokens:     result.Usage.PromptTokens,
-				CompletionTokens: result.Usage.CompletionTokens,
-				TotalTokens:      result.Usage.TotalTokens,
-			},
-		}, nil
-	})
+	return HandleNonStreamingResponse(body, modelUsage, chatModelConfig, OpenAIParserConfig)
 }
 
 func (h *HuaweiCloudModel) ChatStreamlyWithSender(ctx context.Context, modelName string, messages []Message, apiConfig *APIConfig, chatModelConfig *ChatConfig, modelUsage *common.ModelUsage, sender func(*string, *string) error) error {
@@ -325,93 +233,13 @@ func (h *HuaweiCloudModel) ChatStreamlyWithSender(ctx context.Context, modelName
 	huaweiCloudApplyChatConfig(reqBody, modelName, chatModelConfig)
 	if chatModelConfig != nil {
 		chatModelConfig.ToolCallsResult = nil
+		chatModelConfig.UsageResult = nil
 	}
+	reqBody["stream_options"] = map[string]interface{}{"include_usage": true}
 
-	jsonData, err := json.Marshal(reqBody)
-	if err != nil {
-		return fmt.Errorf("failed to marshal request: %w", err)
-	}
-
-	ctx, cancel := context.WithTimeout(ctx, streamCallTimeout)
-	defer cancel()
-
-	req, err := http.NewRequestWithContext(ctx, http.MethodPost, url, bytes.NewBuffer(jsonData))
-	if err != nil {
-		return fmt.Errorf("failed to create request: %w", err)
-	}
-	req.Header.Set("Content-Type", "application/json")
-	req.Header.Set("Authorization", huaweiCloudAuthorization(*apiConfig.ApiKey))
-
-	resp, err := h.baseModel.httpClient.Do(req)
-	if err != nil {
-		return fmt.Errorf("failed to send request: %w", err)
-	}
-	defer resp.Body.Close()
-
-	if resp.StatusCode != http.StatusOK {
-		body, _ := io.ReadAll(resp.Body)
-		return fmt.Errorf("Huawei Cloud stream API error: status %d, body: %s", resp.StatusCode, string(body))
-	}
-
-	sawTerminal := false
-	accumulatedToolCalls := make(map[int]map[string]interface{})
-	done, err := ParseSSEStream[map[string]interface{}](resp.Body, func(event map[string]interface{}) error {
-		if apiErr, ok := event["error"]; ok {
-			return fmt.Errorf("huaweicloud: upstream stream error: %v", apiErr)
-		}
-
-		tokenUsage, found, usageErr := decodeOpenAICompatibleStreamUsage(event)
-		if usageErr != nil {
-			return usageErr
-		}
-		if found {
-			applyStreamUsage(chatModelConfig, modelUsage, tokenUsage)
-		}
-
-		choices, ok := event["choices"].([]interface{})
-		if !ok || len(choices) == 0 {
-			return nil
-		}
-		firstChoice, ok := choices[0].(map[string]interface{})
-		if !ok {
-			return nil
-		}
-		if finishReason, ok := firstChoice["finish_reason"].(string); ok && finishReason != "" {
-			sawTerminal = true
-		}
-		delta, ok := firstChoice["delta"].(map[string]interface{})
-		if !ok {
-			return nil
-		}
-		if accumulateToolCallDeltas(delta, accumulatedToolCalls) {
-			return nil
-		}
-
-		if r, ok := delta["reasoning_content"].(string); ok && r != "" {
-			if err := sender(nil, &r); err != nil {
-				return err
-			}
-		}
-		if content, ok := delta["content"].(string); ok && content != "" {
-			if err := sender(&content, nil); err != nil {
-				return err
-			}
-		}
-		return nil
+	return h.baseModel.doStreamRequest(ctx, url, apiConfig, reqBody, streamCallTimeout, func(body io.ReadCloser) error {
+		return HandleStreamingResponse(body, modelUsage, chatModelConfig, OpenAIParserConfig, sender)
 	})
-	if err != nil {
-		return fmt.Errorf("failed to scan response body: %w", err)
-	}
-	if !done && !sawTerminal {
-		return fmt.Errorf("huaweicloud: stream ended before [DONE] or finish_reason")
-	}
-	setSortedToolCallsResult(chatModelConfig, accumulatedToolCalls)
-
-	endOfStream := "[DONE]"
-	if err := sender(&endOfStream, nil); err != nil {
-		return err
-	}
-	return nil
 }
 
 type huaweiCloudEmbeddingResponse struct {
diff --git a/internal/entity/models/huggingface.go b/internal/entity/models/huggingface.go
index bc3d5145be..974fb2b903 100644
--- a/internal/entity/models/huggingface.go
+++ b/internal/entity/models/huggingface.go
@@ -66,7 +66,6 @@ func (h *HuggingFaceModel) ChatWithMessages(ctx context.Context, modelName strin
 	reqBody := buildRequestBody(chatModelConfig, modelName, messages, false)
 
 	if chatModelConfig != nil {
-
 		if chatModelConfig.Thinking != nil {
 			if *chatModelConfig.Thinking {
 				reqBody["thinking"] = map[string]interface{}{
@@ -80,83 +79,12 @@ func (h *HuggingFaceModel) ChatWithMessages(ctx context.Context, modelName strin
 		}
 	}
 
-	jsonData, err := json.Marshal(reqBody)
+	body, err := h.baseModel.doRequest(ctx, url, apiConfig, reqBody, nonStreamCallTimeout)
 	if err != nil {
-		return nil, fmt.Errorf("failed to marshal request: %w", err)
+		return nil, err
 	}
 
-	ctx, cancel := context.WithTimeout(ctx, nonStreamCallTimeout)
-	defer cancel()
-
-	req, err := http.NewRequestWithContext(ctx, "POST", url, bytes.NewBuffer(jsonData))
-	if err != nil {
-		return nil, fmt.Errorf("failed to create request: %w", err)
-	}
-
-	req.Header.Set("Content-Type", "application/json")
-	req.Header.Set("Authorization", fmt.Sprintf("Bearer %s", *apiConfig.ApiKey))
-
-	resp, err := h.baseModel.httpClient.Do(req)
-	if err != nil {
-		return nil, fmt.Errorf("failed to send request: %w", err)
-	}
-	defer resp.Body.Close()
-
-	body, err := io.ReadAll(resp.Body)
-	if err != nil {
-		return nil, fmt.Errorf("failed to read response: %w", err)
-	}
-
-	if resp.StatusCode != http.StatusOK {
-		return nil, fmt.Errorf("API request failed with status %d: %s", resp.StatusCode, string(body))
-	}
-
-	// Parse response
-	var result map[string]interface{}
-	if err = json.Unmarshal(body, &result); err != nil {
-		return nil, fmt.Errorf("failed to parse response: %w", err)
-	}
-
-	choices, ok := result["choices"].([]interface{})
-	if !ok || len(choices) == 0 {
-		return nil, fmt.Errorf("no choices in response")
-	}
-
-	firstChoice, ok := choices[0].(map[string]interface{})
-	if !ok {
-		return nil, fmt.Errorf("invalid choice format")
-	}
-
-	messageMap, ok := firstChoice["message"].(map[string]interface{})
-	if !ok {
-		return nil, fmt.Errorf("invalid message format")
-	}
-
-	content, ok := messageMap["content"].(string)
-	toolCalls := extractToolCalls(messageMap)
-	if !ok && len(toolCalls) == 0 {
-		return nil, fmt.Errorf("invalid content format")
-	}
-
-	var reasonContent string
-	if chatModelConfig != nil && chatModelConfig.Thinking != nil && *chatModelConfig.Thinking {
-		reasonContent, ok = messageMap["reasoning_content"].(string)
-		if !ok {
-			return nil, fmt.Errorf("invalid content format")
-		}
-		// if first char of reasonContent is \n remove the \n
-		if reasonContent != "" && reasonContent[0] == '\n' {
-			reasonContent = reasonContent[1:]
-		}
-	}
-
-	chatResponse := &ChatResponse{
-		Answer:        &content,
-		ReasonContent: &reasonContent,
-		ToolCalls:     toolCalls,
-	}
-
-	return chatResponse, nil
+	return HandleNonStreamingResponse(body, modelUsage, chatModelConfig, OpenAIParserConfig)
 }
 
 func (h *HuggingFaceModel) ChatStreamlyWithSender(ctx context.Context, modelName string, messages []Message, apiConfig *APIConfig, modelConfig *ChatConfig, modelUsage *common.ModelUsage, sender func(*string, *string) error) error {
@@ -187,77 +115,11 @@ func (h *HuggingFaceModel) ChatStreamlyWithSender(ctx context.Context, modelName
 		}
 	}
 
-	jsonData, err := json.Marshal(reqBody)
-	if err != nil {
-		return fmt.Errorf("failed to marshal request: %w", err)
-	}
+	reqBody["stream_options"] = map[string]interface{}{"include_usage": true}
 
-	ctx, cancel := context.WithTimeout(ctx, streamCallTimeout)
-	defer cancel()
-
-	req, err := http.NewRequestWithContext(ctx, "POST", url, bytes.NewBuffer(jsonData))
-	if err != nil {
-		return fmt.Errorf("failed to create request: %w", err)
-	}
-
-	req.Header.Set("Content-Type", "application/json")
-	req.Header.Set("Authorization", fmt.Sprintf("Bearer %s", *apiConfig.ApiKey))
-
-	resp, err := h.baseModel.httpClient.Do(req)
-	if err != nil {
-		return fmt.Errorf("failed to send request: %w", err)
-	}
-	defer resp.Body.Close()
-
-	if resp.StatusCode != http.StatusOK {
-		body, _ := io.ReadAll(resp.Body)
-		return fmt.Errorf("API request failed with status %d: %s", resp.StatusCode, string(body))
-	}
-
-	accumulatedToolCalls := make(map[int]map[string]any)
-	if _, err := ParseSSEStream[map[string]interface{}](resp.Body, func(event map[string]interface{}) error {
-		common.Info(fmt.Sprintf("%v", event))
-
-		choices, ok := event["choices"].([]interface{})
-		if !ok || len(choices) == 0 {
-			return nil
-		}
-
-		firstChoice, ok := choices[0].(map[string]interface{})
-		if !ok {
-			return nil
-		}
-
-		delta, ok := firstChoice["delta"].(map[string]interface{})
-		if !ok {
-			return nil
-		}
-
-		accumulateToolCallDeltas(delta, accumulatedToolCalls)
-
-		reasoningContent, ok := delta["reasoning_content"].(string)
-		if ok && reasoningContent != "" {
-			if err := sender(nil, &reasoningContent); err != nil {
-				return err
-			}
-		}
-
-		content, ok := delta["content"].(string)
-		if ok && content != "" {
-			if err := sender(&content, nil); err != nil {
-				return err
-			}
-		}
-
-		return nil
-	}); err != nil {
-		return fmt.Errorf("failed to scan response body: %w", err)
-	}
-	setSortedToolCallsResult(modelConfig, accumulatedToolCalls)
-
-	// Send [DONE] marker for OpenAI compatibility
-	endOfStream := "[DONE]"
-	return sender(&endOfStream, nil)
+	return h.baseModel.doStreamRequest(ctx, url, apiConfig, reqBody, streamCallTimeout, func(body io.ReadCloser) error {
+		return HandleStreamingResponse(body, modelUsage, modelConfig, OpenAIParserConfig, sender)
+	})
 }
 
 func (h *HuggingFaceModel) Embed(ctx context.Context, modelName *string, texts []string, apiConfig *APIConfig, embeddingConfig *EmbeddingConfig, modelUsage *common.ModelUsage) ([]EmbeddingData, error) {
diff --git a/internal/entity/models/hunyuan.go b/internal/entity/models/hunyuan.go
index 33d83c2da2..085ac55568 100644
--- a/internal/entity/models/hunyuan.go
+++ b/internal/entity/models/hunyuan.go
@@ -51,32 +51,6 @@ func (h *HunyuanModel) Name() string {
 	return "Tencent Hunyuan"
 }
 
-// HunyuanChatResponse mirrors Tencent Hunyuan's OpenAI-compatible chat response.
-type HunyuanChatResponse struct {
-	ID      string `json:"id"`
-	Object  string `json:"object"`
-	Created int64  `json:"created"`
-	Model   string `json:"model"`
-	Choices []struct {
-		FinishReason string `json:"finish_reason"`
-		Index        int    `json:"index"`
-		Logprobs     any    `json:"logprobs"`
-		Message      struct {
-			Content          string           `json:"content"`
-			Reasoning        string           `json:"reasoning"`
-			ReasoningContent string           `json:"reasoning_content"`
-			Role             string           `json:"role"`
-			ToolCalls        []map[string]any `json:"tool_calls"`
-		} `json:"message"`
-	} `json:"choices"`
-	SystemFingerprint string `json:"system_fingerprint"`
-	Usage             struct {
-		CompletionTokens int `json:"completion_tokens"`
-		PromptTokens     int `json:"prompt_tokens"`
-		TotalTokens      int `json:"total_tokens"`
-	} `json:"usage"`
-}
-
 // HunyuanEmbeddingResponse mirrors Tencent Hunyuan's embeddings response.
 type HunyuanEmbeddingResponse struct {
 	ID     string `json:"id"`
@@ -110,64 +84,12 @@ func (h *HunyuanModel) ChatWithMessages(ctx context.Context, modelName string, m
 	url := fmt.Sprintf("%s/%s", baseURL, h.baseModel.URLSuffix.Chat)
 	reqBody := buildRequestBody(chatModelConfig, modelName, messages, false)
 
-	jsonData, err := json.Marshal(reqBody)
+	body, err := h.baseModel.doRequest(ctx, url, apiConfig, reqBody, nonStreamCallTimeout)
 	if err != nil {
-		return nil, fmt.Errorf("failed to marshal request: %w", err)
+		return nil, err
 	}
 
-	ctx, cancel := context.WithTimeout(ctx, nonStreamCallTimeout)
-	defer cancel()
-
-	req, err := http.NewRequestWithContext(ctx, "POST", url, bytes.NewBuffer(jsonData))
-	if err != nil {
-		return nil, fmt.Errorf("failed to create request: %w", err)
-	}
-	req.Header.Set("Content-Type", "application/json")
-	req.Header.Set("Authorization", fmt.Sprintf("Bearer %s", *apiConfig.ApiKey))
-
-	resp, err := h.baseModel.httpClient.Do(req)
-	if err != nil {
-		return nil, fmt.Errorf("failed to send request: %w", err)
-	}
-	defer resp.Body.Close()
-
-	body, err := io.ReadAll(resp.Body)
-	if err != nil {
-		return nil, fmt.Errorf("failed to read response: %w", err)
-	}
-	if resp.StatusCode != http.StatusOK {
-		return nil, fmt.Errorf("API request failed with status %d: %s", resp.StatusCode, string(body))
-	}
-
-	return parseChatCompletionResponse(body, chatModelConfig, modelUsage, func(body []byte, _ *ChatConfig) (chatResponseParts, error) {
-		var result HunyuanChatResponse
-		if err := json.Unmarshal(body, &result); err != nil {
-			return chatResponseParts{}, fmt.Errorf("failed to parse response: %w", err)
-		}
-		if len(result.Choices) == 0 {
-			return chatResponseParts{}, fmt.Errorf("no choices in response")
-		}
-
-		choice := result.Choices[0]
-		if choice.Message.Content == "" && len(choice.Message.ToolCalls) == 0 {
-			return chatResponseParts{}, fmt.Errorf("response contains neither content nor tool calls")
-		}
-		reasonContent := choice.Message.ReasoningContent
-		if reasonContent == "" {
-			reasonContent = choice.Message.Reasoning
-		}
-		return chatResponseParts{
-			RequestID:     result.ID,
-			Content:       &choice.Message.Content,
-			ReasonContent: &reasonContent,
-			ToolCalls:     choice.Message.ToolCalls,
-			Usage: &TokenUsage{
-				PromptTokens:     result.Usage.PromptTokens,
-				CompletionTokens: result.Usage.CompletionTokens,
-				TotalTokens:      result.Usage.TotalTokens,
-			},
-		}, nil
-	})
+	return HandleNonStreamingResponse(body, modelUsage, chatModelConfig, OpenAIParserConfig)
 }
 
 // ChatStreamlyWithSender opens the SSE chat-completions
@@ -199,94 +121,9 @@ func (h *HunyuanModel) ChatStreamlyWithSender(ctx context.Context, modelName str
 	reqBody := buildRequestBody(chatModelConfig, modelName, messages, true)
 	reqBody["stream_options"] = map[string]any{"include_usage": true}
 
-	jsonData, err := json.Marshal(reqBody)
-	if err != nil {
-		return fmt.Errorf("failed to marshal request: %w", err)
-	}
-
-	ctx, cancel := context.WithTimeout(ctx, streamCallTimeout)
-	defer cancel()
-
-	req, err := http.NewRequestWithContext(ctx, http.MethodPost, url, bytes.NewBuffer(jsonData))
-	if err != nil {
-		return fmt.Errorf("failed to create request: %w", err)
-	}
-	req.Header.Set("Content-Type", "application/json")
-	req.Header.Set("Authorization", fmt.Sprintf("Bearer %s", *apiConfig.ApiKey))
-
-	resp, err := h.baseModel.httpClient.Do(req)
-	if err != nil {
-		return fmt.Errorf("failed to send request: %w", err)
-	}
-	defer resp.Body.Close()
-
-	if resp.StatusCode != http.StatusOK {
-		body, _ := io.ReadAll(resp.Body)
-		return fmt.Errorf("API request failed with status %d: %s", resp.StatusCode, string(body))
-	}
-
-	sawTerminal := false
-	accumulatedToolCalls := make(map[int]map[string]any)
-	done, err := ParseSSEStream[map[string]any](resp.Body, func(event map[string]any) error {
-		common.Info(fmt.Sprintf("%v", event))
-
-		if apiErr, ok := event["error"]; ok {
-			return fmt.Errorf("hunyuan: upstream stream error: %v", apiErr)
-		}
-		tokenUsage, found, usageErr := decodeOpenAICompatibleStreamUsage(event)
-		if usageErr != nil {
-			return usageErr
-		}
-		if found {
-			applyStreamUsage(chatModelConfig, modelUsage, tokenUsage)
-		}
-		choices, ok := event["choices"].([]any)
-		if !ok || len(choices) == 0 {
-			return nil
-		}
-
-		choice, ok := choices[0].(map[string]any)
-		if !ok {
-			return nil
-		}
-		if finishReason, ok := choice["finish_reason"].(string); ok && finishReason != "" {
-			sawTerminal = true
-		}
-		delta, ok := choice["delta"].(map[string]any)
-		if !ok {
-			return nil
-		}
-		accumulateToolCallDeltas(delta, accumulatedToolCalls)
-		if reasoning, ok := delta["reasoning_content"].(string); ok && reasoning != "" {
-			if err := sender(nil, &reasoning); err != nil {
-				return err
-			}
-		}
-		if reasoning, ok := delta["reasoning"].(string); ok && reasoning != "" {
-			if err := sender(nil, &reasoning); err != nil {
-				return err
-			}
-		}
-		if content, ok := delta["content"].(string); ok && content != "" {
-			if err := sender(&content, nil); err != nil {
-				return err
-			}
-		}
-		return nil
+	return h.baseModel.doStreamRequest(ctx, url, apiConfig, reqBody, streamCallTimeout, func(body io.ReadCloser) error {
+		return HandleStreamingResponse(body, modelUsage, chatModelConfig, OpenAIParserConfig, sender)
 	})
-	if err != nil {
-		return fmt.Errorf("failed to scan response body: %w", err)
-	}
-	setSortedToolCallsResult(chatModelConfig, accumulatedToolCalls)
-	if !done && !sawTerminal {
-		return fmt.Errorf("hunyuan: stream ended before [DONE] or finish_reason")
-	}
-
-	endOfStream := "[DONE]"
-	if err := sender(&endOfStream, nil); err != nil {
-		return err
-	}
-	return nil
 }
 
 func (h *HunyuanModel) ListModels(ctx context.Context, apiConfig *APIConfig) ([]ListModelResponse, error) {
diff --git a/internal/entity/models/hunyuan_test.go b/internal/entity/models/hunyuan_test.go
index dc3ba54e86..cdd2de77df 100644
--- a/internal/entity/models/hunyuan_test.go
+++ b/internal/entity/models/hunyuan_test.go
@@ -360,7 +360,8 @@ func TestHunyuanStreamFailsWithoutTerminal(t *testing.T) {
 func TestHunyuanStreamAcceptsTerminalWithoutDelta(t *testing.T) {
 	withSSRFBypass(t)
 	srv := newHunyuanSSEServer(t, "/chat/completions",
-		`data: {"choices":[{"finish_reason":"stop"}]}`+"\n\n",
+		`data: {"choices":[{"finish_reason":"stop"}]}`+"\n\n"+
+			`data: [DONE]`+"\n\n",
 	)
 	defer srv.Close()
 
diff --git a/internal/entity/models/jiekouai.go b/internal/entity/models/jiekouai.go
index 156f69c99e..8e6b035a67 100644
--- a/internal/entity/models/jiekouai.go
+++ b/internal/entity/models/jiekouai.go
@@ -62,32 +62,6 @@ func validateJieKouAIModelName(modelName *string) (string, error) {
 	return strings.TrimSpace(*modelName), nil
 }
 
-// JieKouAIChatResponse mirrors Jiekou.AI's OpenAI-compatible chat response.
-type JieKouAIChatResponse struct {
-	ID      string `json:"id"`
-	Object  string `json:"object"`
-	Created int64  `json:"created"`
-	Model   string `json:"model"`
-	Choices []struct {
-		FinishReason string `json:"finish_reason"`
-		Index        int    `json:"index"`
-		Logprobs     any    `json:"logprobs"`
-		Message      struct {
-			Content          string           `json:"content"`
-			Reasoning        string           `json:"reasoning"`
-			ReasoningContent string           `json:"reasoning_content"`
-			Role             string           `json:"role"`
-			ToolCalls        []map[string]any `json:"tool_calls"`
-		} `json:"message"`
-	} `json:"choices"`
-	SystemFingerprint string `json:"system_fingerprint"`
-	Usage             struct {
-		CompletionTokens int `json:"completion_tokens"`
-		PromptTokens     int `json:"prompt_tokens"`
-		TotalTokens      int `json:"total_tokens"`
-	} `json:"usage"`
-}
-
 // JieKouAIEmbeddingResponse mirrors Jiekou.AI's embeddings response.
 type JieKouAIEmbeddingResponse struct {
 	ID     string `json:"id"`
@@ -127,7 +101,6 @@ func (j *JieKouAIModel) ChatWithMessages(ctx context.Context, modelName string,
 	if err := j.baseModel.APIConfigCheck(apiConfig); err != nil {
 		return nil, err
 	}
-	apiKey := strings.TrimSpace(*apiConfig.ApiKey)
 	if modelName = strings.TrimSpace(modelName); modelName == "" {
 		return nil, fmt.Errorf("model name is required")
 	}
@@ -151,78 +124,18 @@ func (j *JieKouAIModel) ChatWithMessages(ctx context.Context, modelName string,
 		}
 	}
 
-	jsonData, err := json.Marshal(reqBody)
+	body, err := j.baseModel.doRequest(ctx, url, apiConfig, reqBody, nonStreamCallTimeout)
 	if err != nil {
-		return nil, fmt.Errorf("failed to marshal request: %w", err)
+		return nil, err
 	}
 
-	ctx, cancel := context.WithTimeout(ctx, nonStreamCallTimeout)
-	defer cancel()
-
-	req, err := http.NewRequestWithContext(ctx, http.MethodPost, url, bytes.NewBuffer(jsonData))
-	if err != nil {
-		return nil, fmt.Errorf("failed to create request: %w", err)
-	}
-
-	req.Header.Set("Content-Type", "application/json")
-	req.Header.Set("Authorization", fmt.Sprintf("Bearer %s", apiKey))
-	req.Header.Set("Accept", "application/json")
-
-	resp, err := j.baseModel.httpClient.Do(req)
-	if err != nil {
-		return nil, fmt.Errorf("failed to send request: %w", err)
-	}
-	defer resp.Body.Close()
-
-	body, err := io.ReadAll(resp.Body)
-	if err != nil {
-		return nil, fmt.Errorf("failed to read response: %w", err)
-	}
-
-	if resp.StatusCode != http.StatusOK {
-		return nil, fmt.Errorf("API request failed with status %d: %s", resp.StatusCode, string(body))
-	}
-
-	return parseChatCompletionResponse(body, chatModelConfig, modelUsage, func(body []byte, chatConfig *ChatConfig) (chatResponseParts, error) {
-		var result JieKouAIChatResponse
-		if err := json.Unmarshal(body, &result); err != nil {
-			return chatResponseParts{}, fmt.Errorf("failed to parse response: %w", err)
-		}
-		if len(result.Choices) == 0 {
-			return chatResponseParts{}, fmt.Errorf("no choices in response")
-		}
-
-		choice := result.Choices[0]
-		if choice.Message.Content == "" && len(choice.Message.ToolCalls) == 0 {
-			return chatResponseParts{}, fmt.Errorf("response contains neither content nor tool calls")
-		}
-		reasonContent := ""
-		if chatConfig != nil && chatConfig.Thinking != nil && *chatConfig.Thinking {
-			reasonContent = choice.Message.ReasoningContent
-			if reasonContent == "" {
-				reasonContent = choice.Message.Reasoning
-			}
-			reasonContent = strings.TrimPrefix(reasonContent, "\n")
-		}
-		return chatResponseParts{
-			RequestID:     result.ID,
-			Content:       &choice.Message.Content,
-			ReasonContent: &reasonContent,
-			ToolCalls:     choice.Message.ToolCalls,
-			Usage: &TokenUsage{
-				PromptTokens:     result.Usage.PromptTokens,
-				CompletionTokens: result.Usage.CompletionTokens,
-				TotalTokens:      result.Usage.TotalTokens,
-			},
-		}, nil
-	})
+	return HandleNonStreamingResponse(body, modelUsage, chatModelConfig, OpenAIParserConfig)
 }
 
 func (j *JieKouAIModel) ChatStreamlyWithSender(ctx context.Context, modelName string, messages []Message, apiConfig *APIConfig, modelConfig *ChatConfig, modelUsage *common.ModelUsage, sender func(*string, *string) error) error {
 	if err := j.baseModel.APIConfigCheck(apiConfig); err != nil {
 		return err
 	}
-	apiKey := strings.TrimSpace(*apiConfig.ApiKey)
 	if modelName = strings.TrimSpace(modelName); modelName == "" {
 		return fmt.Errorf("model name is required")
 	}
@@ -253,96 +166,9 @@ func (j *JieKouAIModel) ChatStreamlyWithSender(ctx context.Context, modelName st
 		}
 	}
 
-	jsonData, err := json.Marshal(reqBody)
-	if err != nil {
-		return fmt.Errorf("failed to marshal request: %w", err)
-	}
-
-	ctx, cancel := context.WithTimeout(ctx, streamCallTimeout)
-	defer cancel()
-
-	req, err := http.NewRequestWithContext(ctx, http.MethodPost, url, bytes.NewBuffer(jsonData))
-	if err != nil {
-		return fmt.Errorf("failed to create request: %w", err)
-	}
-
-	req.Header.Set("Content-Type", "application/json")
-	req.Header.Set("Authorization", fmt.Sprintf("Bearer %s", apiKey))
-	req.Header.Set("Accept", "text/event-stream")
-
-	resp, err := j.baseModel.httpClient.Do(req)
-	if err != nil {
-		return fmt.Errorf("failed to send request: %w", err)
-	}
-	defer resp.Body.Close()
-
-	if resp.StatusCode != http.StatusOK {
-		body, _ := io.ReadAll(resp.Body)
-		return fmt.Errorf("API request failed with status %d: %s", resp.StatusCode, string(body))
-	}
-
-	sawTerminal := false
-	accumulatedToolCalls := make(map[int]map[string]any)
-	done, err := ParseSSEStream[map[string]any](resp.Body, func(event map[string]any) error {
-		common.Info(fmt.Sprintf("%v", event))
-
-		tokenUsage, found, usageErr := decodeOpenAICompatibleStreamUsage(event)
-		if usageErr != nil {
-			return usageErr
-		}
-		if found {
-			applyStreamUsage(modelConfig, modelUsage, tokenUsage)
-		}
-		choices, ok := event["choices"].([]any)
-		if !ok || len(choices) == 0 {
-			return nil
-		}
-
-		choice, ok := choices[0].(map[string]any)
-		if !ok {
-			return nil
-		}
-		if finishReason, ok := choice["finish_reason"].(string); ok && finishReason != "" {
-			sawTerminal = true
-		}
-		delta, ok := choice["delta"].(map[string]any)
-		if !ok {
-			return nil
-		}
-
-		accumulateToolCallDeltas(delta, accumulatedToolCalls)
-
-		reasoningContent, ok := delta["reasoning_content"].(string)
-		if ok && reasoningContent != "" {
-			if err := sender(nil, &reasoningContent); err != nil {
-				return err
-			}
-		}
-
-		content, ok := delta["content"].(string)
-		if ok && content != "" {
-			if err := sender(&content, nil); err != nil {
-				return err
-			}
-		}
-
-		return nil
+	return j.baseModel.doStreamRequest(ctx, url, apiConfig, reqBody, streamCallTimeout, func(body io.ReadCloser) error {
+		return HandleStreamingResponse(body, modelUsage, modelConfig, OpenAIParserConfig, sender)
 	})
-	if err != nil {
-		return fmt.Errorf("failed to scan response body: %w", err)
-	}
-	if !done && !sawTerminal {
-		return fmt.Errorf("jiekouai: stream ended before [DONE] or finish_reason")
-	}
-	setSortedToolCallsResult(modelConfig, accumulatedToolCalls)
-
-	// Send [DONE] marker for OpenAI compatibility
-	endOfStream := "[DONE]"
-	if err = sender(&endOfStream, nil); err != nil {
-		return err
-	}
-
-	return nil
 }
 
 func (j *JieKouAIModel) Embed(ctx context.Context, modelName *string, texts []string, apiConfig *APIConfig, embeddingConfig *EmbeddingConfig, modelUsage *common.ModelUsage) ([]EmbeddingData, error) {
diff --git a/internal/entity/models/jina.go b/internal/entity/models/jina.go
index 241e27e909..f418ffdbb5 100644
--- a/internal/entity/models/jina.go
+++ b/internal/entity/models/jina.go
@@ -55,35 +55,6 @@ func (j *JinaModel) Name() string {
 	return "jina"
 }
 
-// JinaChatResponse mirrors Jina's OpenAI-compatible chat response.
-type JinaChatResponse struct {
-	ID      string `json:"id"`
-	Object  string `json:"object"`
-	Created int64  `json:"created"`
-	Model   string `json:"model"`
-	Choices []struct {
-		FinishReason string `json:"finish_reason"`
-		Index        int    `json:"index"`
-		Logprobs     any    `json:"logprobs"`
-		Message      struct {
-			Content          string           `json:"content"`
-			Reasoning        string           `json:"reasoning"`
-			ReasoningContent string           `json:"reasoning_content"`
-			Role             string           `json:"role"`
-			ToolCalls        []map[string]any `json:"tool_calls"`
-		} `json:"message"`
-	} `json:"choices"`
-	SystemFingerprint string `json:"system_fingerprint"`
-	Usage             struct {
-		CompletionTokens    int `json:"completion_tokens"`
-		PromptTokens        int `json:"prompt_tokens"`
-		TotalTokens         int `json:"total_tokens"`
-		PromptTokensDetails struct {
-			CachedTokens int `json:"cached_tokens"`
-		} `json:"prompt_tokens_details"`
-	} `json:"usage"`
-}
-
 // JinaEmbeddingResponse mirrors Jina's embeddings response. Embeddings is
 // populated by multivector models such as jina-embeddings-v4.
 type JinaEmbeddingResponse struct {
@@ -139,72 +110,55 @@ func (j *JinaModel) ChatWithMessages(ctx context.Context, modelName string, mess
 
 	reqBody := buildRequestBody(chatModelConfig, modelName, messages, false)
 
-	jsonData, err := json.Marshal(reqBody)
-	if err != nil {
-		return nil, fmt.Errorf("failed to marshal request: %w", err)
-	}
-
-	ctx, cancel := context.WithTimeout(ctx, nonStreamCallTimeout)
-	defer cancel()
-
-	req, err := http.NewRequestWithContext(ctx, "POST", url, bytes.NewBuffer(jsonData))
-	if err != nil {
-		return nil, fmt.Errorf("failed to create request: %w", err)
-	}
-
-	req.Header.Set("Content-Type", "application/json")
-	req.Header.Set("Authorization", fmt.Sprintf("Bearer %s", *apiConfig.ApiKey))
-
-	resp, err := j.baseModel.httpClient.Do(req)
-	if err != nil {
-		return nil, fmt.Errorf("failed to send request: %w", err)
-	}
-	defer resp.Body.Close()
-
-	body, err := io.ReadAll(resp.Body)
-	if err != nil {
-		return nil, fmt.Errorf("failed to read response: %w", err)
-	}
-
-	if resp.StatusCode != http.StatusOK {
-		return nil, fmt.Errorf("Jina chat API error: status %d, body: %s", resp.StatusCode, string(body))
-	}
-
-	return parseChatCompletionResponse(body, chatModelConfig, modelUsage, func(body []byte, _ *ChatConfig) (chatResponseParts, error) {
-		var result JinaChatResponse
-		if err := json.Unmarshal(body, &result); err != nil {
-			return chatResponseParts{}, fmt.Errorf("failed to parse response: %w", err)
-		}
-		if len(result.Choices) == 0 {
-			return chatResponseParts{}, fmt.Errorf("no choices in response")
+	if chatModelConfig != nil {
+		if chatModelConfig.Thinking != nil {
+			reqBody["enable_thinking"] = *chatModelConfig.Thinking
 		}
+	}
 
-		choice := result.Choices[0]
-		if choice.Message.Content == "" && len(choice.Message.ToolCalls) == 0 {
-			return chatResponseParts{}, fmt.Errorf("response contains neither content nor tool calls")
-		}
-		reasonContent := choice.Message.ReasoningContent
-		if reasonContent == "" {
-			reasonContent = choice.Message.Reasoning
-		}
-		reasonContent = strings.TrimPrefix(reasonContent, "\n")
-		return chatResponseParts{
-			RequestID:     result.ID,
-			Content:       &choice.Message.Content,
-			ReasonContent: &reasonContent,
-			ToolCalls:     choice.Message.ToolCalls,
-			Usage: &TokenUsage{
-				PromptTokens:     result.Usage.PromptTokens,
-				CompletionTokens: result.Usage.CompletionTokens,
-				TotalTokens:      result.Usage.TotalTokens,
-			},
-		}, nil
-	})
+	body, err := j.baseModel.doRequest(ctx, url, apiConfig, reqBody, nonStreamCallTimeout)
+	if err != nil {
+		return nil, err
+	}
+
+	return HandleNonStreamingResponse(body, modelUsage, chatModelConfig, OpenAIParserConfig)
 }
 
-func (j *JinaModel) ChatStreamlyWithSender(ctx context.Context, modelName string, messages []Message, apiConfig *APIConfig, modelConfig *ChatConfig, modelUsage *common.ModelUsage, sender func(*string, *string) error) error {
-	// Jina's public API does not expose a streaming chat-completions endpoint.
-	return fmt.Errorf("jina does not implement ChatStreamlyWithSender(not available for now)")
+func (j *JinaModel) ChatStreamlyWithSender(ctx context.Context, modelName string, messages []Message, apiConfig *APIConfig, chatModelConfig *ChatConfig, modelUsage *common.ModelUsage, sender func(*string, *string) error) error {
+	if err := j.baseModel.APIConfigCheck(apiConfig); err != nil {
+		return err
+	}
+	if modelName == "" {
+		return fmt.Errorf("model name is required")
+	}
+	if len(messages) == 0 {
+		return fmt.Errorf("messages is empty")
+	}
+	if err := validateStreamConfig(chatModelConfig); err != nil {
+		return err
+	}
+
+	baseURL, err := j.baseModel.GetBaseURL(apiConfig)
+	if err != nil {
+		return err
+	}
+	baseURL = strings.TrimSuffix(baseURL, "/")
+	url := fmt.Sprintf("%s/%s", baseURL, j.baseModel.URLSuffix.Chat)
+
+	reqBody := buildRequestBody(chatModelConfig, modelName, messages, true)
+	reqBody["stream_options"] = map[string]interface{}{"include_usage": true}
+
+	if chatModelConfig != nil {
+		chatModelConfig.ToolCallsResult = nil
+		chatModelConfig.UsageResult = nil
+		if chatModelConfig.Thinking != nil {
+			reqBody["enable_thinking"] = *chatModelConfig.Thinking
+		}
+	}
+
+	return j.baseModel.doStreamRequest(ctx, url, apiConfig, reqBody, streamCallTimeout, func(body io.ReadCloser) error {
+		return HandleStreamingResponse(body, modelUsage, chatModelConfig, OpenAIParserConfig, sender)
+	})
 }
 
 func (j *JinaModel) Embed(ctx context.Context, modelName *string, texts []string, apiConfig *APIConfig, embeddingConfig *EmbeddingConfig, modelUsage *common.ModelUsage) ([]EmbeddingData, error) {
diff --git a/internal/entity/models/jina_test.go b/internal/entity/models/jina_test.go
index f2648c50d7..9528ddc8cf 100644
--- a/internal/entity/models/jina_test.go
+++ b/internal/entity/models/jina_test.go
@@ -227,23 +227,6 @@ func TestJinaChatValidation(t *testing.T) {
 	}
 }
 
-func TestJinaChatStreamIsNotSupported(t *testing.T) {
-	withSSRFBypass(t)
-	apiKey := "test-key"
-	err := newJinaForTest("http://unused").ChatStreamlyWithSender(
-		t.Context(),
-		"jina-vlm",
-		[]Message{{Role: "user", Content: "x"}},
-		&APIConfig{ApiKey: &apiKey},
-		nil,
-		nil,
-		func(*string, *string) error { return nil },
-	)
-	if err == nil || !strings.Contains(err.Error(), "ChatStreamlyWithSender") {
-		t.Fatalf("expected unsupported streaming error, got %v", err)
-	}
-}
-
 func TestJinaEmbedMeanPoolsMultivectorResponse(t *testing.T) {
 	withSSRFBypass(t)
 	srv := newJinaServer(t, "/embeddings", func(t *testing.T, _ map[string]interface{}, w http.ResponseWriter) {
diff --git a/internal/entity/models/lmstudio.go b/internal/entity/models/lmstudio.go
index 29b0d5ffaa..af815a0678 100644
--- a/internal/entity/models/lmstudio.go
+++ b/internal/entity/models/lmstudio.go
@@ -91,84 +91,12 @@ func (l *LmStudioModel) ChatWithMessages(ctx context.Context, modelName string,
 		}
 	}
 
-	jsonData, err := json.Marshal(reqBody)
+	body, err := l.baseModel.doRequest(ctx, url, apiConfig, reqBody, nonStreamCallTimeout)
 	if err != nil {
-		return nil, fmt.Errorf("failed to marshal request: %w", err)
+		return nil, err
 	}
 
-	ctx, cancel := context.WithTimeout(ctx, nonStreamCallTimeout)
-	defer cancel()
-
-	req, err := http.NewRequestWithContext(ctx, "POST", url, bytes.NewBuffer(jsonData))
-	if err != nil {
-		return nil, fmt.Errorf("failed to create request: %w", err)
-	}
-
-	req.Header.Set("Content-Type", "application/json")
-	if auth := BearerAuth(apiConfig); auth != "" {
-		req.Header.Set("Authorization", auth)
-	}
-
-	resp, err := l.baseModel.httpClient.Do(req)
-	if err != nil {
-		return nil, fmt.Errorf("failed to send request: %w", err)
-	}
-	defer resp.Body.Close()
-
-	body, err := io.ReadAll(resp.Body)
-	if err != nil {
-		return nil, fmt.Errorf("failed to read response: %w", err)
-	}
-
-	if resp.StatusCode != http.StatusOK {
-		return nil, fmt.Errorf("API request failed with status %d: %s :%s", resp.StatusCode, string(body), messages[0].Content)
-	}
-
-	// Parse response
-	var result map[string]interface{}
-	if err = json.Unmarshal(body, &result); err != nil {
-		return nil, fmt.Errorf("failed to parse response: %w", err)
-	}
-
-	choices, ok := result["choices"].([]interface{})
-	if !ok || len(choices) == 0 {
-		return nil, fmt.Errorf("no choices in response")
-	}
-
-	firstChoice, ok := choices[0].(map[string]interface{})
-	if !ok {
-		return nil, fmt.Errorf("invalid choice format")
-	}
-
-	messageMap, ok := firstChoice["message"].(map[string]interface{})
-	if !ok {
-		return nil, fmt.Errorf("invalid message format")
-	}
-
-	content, ok := messageMap["content"].(string)
-	toolCalls := extractToolCalls(messageMap)
-	if !ok && len(toolCalls) == 0 {
-		return nil, fmt.Errorf("invalid content format")
-	}
-
-	var reasonContent string
-	if chatModelConfig != nil && chatModelConfig.Thinking != nil && *chatModelConfig.Thinking {
-		reasonContent, ok = messageMap["reasoning_content"].(string)
-		if !ok {
-			return nil, fmt.Errorf("invalid content format")
-		}
-		if reasonContent != "" && reasonContent[0] == '\n' {
-			reasonContent = reasonContent[1:]
-		}
-	}
-
-	chatResponse := &ChatResponse{
-		Answer:        &content,
-		ReasonContent: &reasonContent,
-		ToolCalls:     toolCalls,
-	}
-
-	return chatResponse, nil
+	return HandleNonStreamingResponse(body, modelUsage, chatModelConfig, OpenAIParserConfig)
 }
 
 // ChatStreamlyWithSender sends messages and streams response via sender function (best performance, no channel)
@@ -205,84 +133,11 @@ func (l *LmStudioModel) ChatStreamlyWithSender(ctx context.Context, modelName st
 		}
 	}
 
-	jsonData, err := json.Marshal(reqBody)
-	if err != nil {
-		return fmt.Errorf("failed to marshal request: %w", err)
-	}
+	reqBody["stream_options"] = map[string]any{"include_usage": true}
 
-	ctx, cancel := context.WithTimeout(ctx, streamCallTimeout)
-	defer cancel()
-
-	req, err := http.NewRequestWithContext(ctx, "POST", url, bytes.NewBuffer(jsonData))
-	if err != nil {
-		return fmt.Errorf("failed to create request: %w", err)
-	}
-
-	req.Header.Set("Content-Type", "application/json")
-	if auth := BearerAuth(apiConfig); auth != "" {
-		req.Header.Set("Authorization", auth)
-	}
-
-	resp, err := l.baseModel.httpClient.Do(req)
-	if err != nil {
-		return fmt.Errorf("failed to send request: %w", err)
-	}
-	defer resp.Body.Close()
-
-	if resp.StatusCode != http.StatusOK {
-		body, _ := io.ReadAll(resp.Body)
-		return fmt.Errorf("API request failed with status %d: %s", resp.StatusCode, string(body))
-	}
-
-	// SSE parsing: read line by line
-	accumulatedToolCalls := make(map[int]map[string]any)
-	if _, err := ParseSSEStream[map[string]interface{}](resp.Body, func(event map[string]interface{}) error {
-		common.Info(fmt.Sprintf("%v", event))
-
-		choices, ok := event["choices"].([]interface{})
-		if !ok || len(choices) == 0 {
-			return nil
-		}
-
-		firstChoice, ok := choices[0].(map[string]interface{})
-		if !ok {
-			return nil
-		}
-
-		delta, ok := firstChoice["delta"].(map[string]interface{})
-		if !ok {
-			return nil
-		}
-
-		accumulateToolCallDeltas(delta, accumulatedToolCalls)
-
-		reasoningContent, ok := delta["reasoning_content"].(string)
-		if ok && reasoningContent != "" {
-			if err := sender(nil, &reasoningContent); err != nil {
-				return err
-			}
-		}
-
-		content, ok := delta["content"].(string)
-		if ok && content != "" {
-			if err := sender(&content, nil); err != nil {
-				return err
-			}
-		}
-
-		return nil
-	}); err != nil {
-		return fmt.Errorf("failed to scan response body: %w", err)
-	}
-	setSortedToolCallsResult(modelConfig, accumulatedToolCalls)
-
-	// Send [DONE] marker for OpenAI compatibility
-	endOfStream := "[DONE]"
-	if err = sender(&endOfStream, nil); err != nil {
-		return err
-	}
-
-	return nil
+	return l.baseModel.doStreamRequest(ctx, url, apiConfig, reqBody, streamCallTimeout, func(body io.ReadCloser) error {
+		return HandleStreamingResponse(body, modelUsage, modelConfig, OpenAIParserConfig, sender)
+	})
 }
 
 func (l *LmStudioModel) Embed(ctx context.Context, modelName *string, texts []string, apiConfig *APIConfig, embeddingConfig *EmbeddingConfig, modelUsage *common.ModelUsage) ([]EmbeddingData, error) {
diff --git a/internal/entity/models/localai.go b/internal/entity/models/localai.go
index a2e16e4800..442a67d31d 100644
--- a/internal/entity/models/localai.go
+++ b/internal/entity/models/localai.go
@@ -25,13 +25,8 @@ import (
 	"net/http"
 	"ragflow/internal/common"
 	"strings"
-	"sync"
-	"time"
 )
 
-// localAIStreamIdleTimeout bounds how long ChatStreamlyWithSender
-var localAIStreamIdleTimeout = 60 * time.Second
-
 // LocalAIModel implements ModelDriver for LocalAI
 type LocalAIModel struct {
 	baseModel BaseModel
@@ -57,17 +52,6 @@ func (l *LocalAIModel) Name() string {
 	return "LocalAI"
 }
 
-var localAIReasoningFields = []string{"reasoning_content", "reasoning", "thinking"}
-
-func extractLocalAIReasoning(m map[string]interface{}) string {
-	for _, k := range localAIReasoningFields {
-		if v, ok := m[k].(string); ok && v != "" {
-			return v
-		}
-	}
-	return ""
-}
-
 func addLocalAIReasoningRequestParams(reqBody map[string]interface{}, cfg *ChatConfig) {
 	if cfg == nil {
 		return
@@ -100,74 +84,12 @@ func (l *LocalAIModel) ChatWithMessages(ctx context.Context, modelName string, m
 		addLocalAIReasoningRequestParams(reqBody, chatModelConfig)
 	}
 
-	jsonData, err := json.Marshal(reqBody)
+	body, err := l.baseModel.doRequest(ctx, url, apiConfig, reqBody, nonStreamCallTimeout)
 	if err != nil {
-		return nil, fmt.Errorf("failed to marshal request: %w", err)
+		return nil, err
 	}
 
-	ctx, cancel := context.WithTimeout(ctx, nonStreamCallTimeout)
-	defer cancel()
-
-	req, err := http.NewRequestWithContext(ctx, "POST", url, bytes.NewBuffer(jsonData))
-	if err != nil {
-		return nil, fmt.Errorf("failed to create request: %w", err)
-	}
-
-	req.Header.Set("Content-Type", "application/json")
-	if auth := BearerAuth(apiConfig); auth != "" {
-		req.Header.Set("Authorization", auth)
-	}
-
-	resp, err := l.baseModel.httpClient.Do(req)
-	if err != nil {
-		return nil, fmt.Errorf("failed to send request: %w", err)
-	}
-	defer resp.Body.Close()
-
-	body, err := io.ReadAll(resp.Body)
-	if err != nil {
-		return nil, fmt.Errorf("failed to read response: %w", err)
-	}
-
-	if resp.StatusCode != http.StatusOK {
-		return nil, fmt.Errorf("API request failed with status %d: %s", resp.StatusCode, string(body))
-	}
-
-	var result map[string]interface{}
-	if err = json.Unmarshal(body, &result); err != nil {
-		return nil, fmt.Errorf("failed to parse response: %w", err)
-	}
-
-	choices, ok := result["choices"].([]interface{})
-	if !ok || len(choices) == 0 {
-		return nil, fmt.Errorf("no choices in response")
-	}
-
-	firstChoice, ok := choices[0].(map[string]interface{})
-	if !ok {
-		return nil, fmt.Errorf("invalid choice format")
-	}
-
-	messageMap, ok := firstChoice["message"].(map[string]interface{})
-	if !ok {
-		return nil, fmt.Errorf("invalid message format")
-	}
-
-	content, ok := messageMap["content"].(string)
-	toolCalls := extractToolCalls(messageMap)
-	if !ok && len(toolCalls) == 0 {
-		return nil, fmt.Errorf("invalid content format")
-	}
-
-	// Pull the chain-of-thought from whichever field the upstream model
-	// used. See localAIReasoningFields for the priority order.
-	reasonContent := extractLocalAIReasoning(messageMap)
-
-	return &ChatResponse{
-		Answer:        &content,
-		ReasonContent: &reasonContent,
-		ToolCalls:     toolCalls,
-	}, nil
+	return HandleNonStreamingResponse(body, modelUsage, chatModelConfig, OpenAIParserConfig)
 }
 
 // ChatStreamlyWithSender sends messages and streams the response via the
@@ -197,118 +119,13 @@ func (l *LocalAIModel) ChatStreamlyWithSender(ctx context.Context, modelName str
 		addLocalAIReasoningRequestParams(reqBody, chatModelConfig)
 	}
 
-	jsonData, err := json.Marshal(reqBody)
-	if err != nil {
-		return fmt.Errorf("failed to marshal request: %w", err)
+	reqBody["stream_options"] = map[string]interface{}{
+		"include_usage": true,
 	}
 
-	ctx, cancel := context.WithCancel(context.Background())
-	defer cancel()
-
-	req, err := http.NewRequestWithContext(ctx, "POST", url, bytes.NewBuffer(jsonData))
-	if err != nil {
-		return fmt.Errorf("failed to create request: %w", err)
-	}
-
-	req.Header.Set("Content-Type", "application/json")
-	if auth := BearerAuth(apiConfig); auth != "" {
-		req.Header.Set("Authorization", auth)
-	}
-
-	resp, err := l.baseModel.httpClient.Do(req)
-	if err != nil {
-		return fmt.Errorf("failed to send request: %w", err)
-	}
-	defer resp.Body.Close()
-
-	if resp.StatusCode != http.StatusOK {
-		body, _ := io.ReadAll(resp.Body)
-		return fmt.Errorf("API request failed with status %d: %s", resp.StatusCode, string(body))
-	}
-
-	lastActive := time.Now()
-	var lastActiveMu sync.Mutex
-	done := make(chan struct{})
-	defer close(done)
-	go func() {
-		ticker := time.NewTicker(localAIStreamIdleTimeout / 4)
-		defer ticker.Stop()
-		for {
-			select {
-			case <-done:
-				return
-			case now := <-ticker.C:
-				lastActiveMu.Lock()
-				idle := now.Sub(lastActive)
-				lastActiveMu.Unlock()
-				if idle >= localAIStreamIdleTimeout {
-					cancel()
-					return
-				}
-			}
-		}
-	}()
-
-	sawTerminal := false
-	accumulatedToolCalls := make(map[int]map[string]any)
-	streamDone, err := ParseSSEStream[map[string]interface{}](resp.Body, func(event map[string]interface{}) error {
-		lastActiveMu.Lock()
-		lastActive = time.Now()
-		lastActiveMu.Unlock()
-
-		choices, ok := event["choices"].([]interface{})
-		if !ok || len(choices) == 0 {
-			return nil
-		}
-
-		firstChoice, ok := choices[0].(map[string]interface{})
-		if !ok {
-			return nil
-		}
-
-		delta, ok := firstChoice["delta"].(map[string]interface{})
-		if !ok {
-			return nil
-		}
-
-		accumulateToolCallDeltas(delta, accumulatedToolCalls)
-
-		if reasoning := extractLocalAIReasoning(delta); reasoning != "" {
-			if err := sender(nil, &reasoning); err != nil {
-				return err
-			}
-		}
-
-		content, ok := delta["content"].(string)
-		if ok && content != "" {
-			if err := sender(&content, nil); err != nil {
-				return err
-			}
-		}
-
-		finishReason, ok := firstChoice["finish_reason"].(string)
-		if ok && finishReason != "" {
-			sawTerminal = true
-		}
-		return nil
+	return l.baseModel.doStreamRequest(ctx, url, apiConfig, reqBody, streamCallTimeout, func(body io.ReadCloser) error {
+		return HandleStreamingResponse(body, modelUsage, chatModelConfig, OpenAIParserConfig, sender)
 	})
-	if err != nil {
-		if ctx.Err() != nil {
-			return fmt.Errorf("localai: stream idle for more than %s, aborted", localAIStreamIdleTimeout)
-		}
-		return fmt.Errorf("failed to scan response body: %w", err)
-	}
-	setSortedToolCallsResult(chatModelConfig, accumulatedToolCalls)
-	if !streamDone && !sawTerminal {
-		return fmt.Errorf("localai: stream ended before [DONE] or finish_reason")
-	}
-
-	endOfStream := "[DONE]"
-	if err := sender(&endOfStream, nil); err != nil {
-		return err
-	}
-
-	return nil
 }
 
 type localAIEmbeddingData struct {
diff --git a/internal/entity/models/localai_test.go b/internal/entity/models/localai_test.go
index b4d03d6544..48ea8af3e5 100644
--- a/internal/entity/models/localai_test.go
+++ b/internal/entity/models/localai_test.go
@@ -6,9 +6,7 @@ import (
 	"net/http"
 	"net/http/httptest"
 	"strings"
-	"sync"
 	"testing"
-	"time"
 )
 
 func newLocalAIForTest(baseURL string) *LocalAIModel {
@@ -23,18 +21,6 @@ func newLocalAIForTest(baseURL string) *LocalAIModel {
 	)
 }
 
-// withLocalAIIdleTimeout swaps the package-level idle timeout for the
-// duration of the test. Tests that exercise the stall watchdog use a
-// sub-second value so they finish quickly.
-func withLocalAIIdleTimeout(t *testing.T, d time.Duration) {
-	t.Helper()
-	original := localAIStreamIdleTimeout
-	localAIStreamIdleTimeout = d
-	t.Cleanup(func() {
-		localAIStreamIdleTimeout = original
-	})
-}
-
 func TestLocalAIName(t *testing.T) {
 	l := newLocalAIForTest("http://unused")
 	if got := l.Name(); got != "LocalAI" {
@@ -42,114 +28,6 @@ func TestLocalAIName(t *testing.T) {
 	}
 }
 
-func TestLocalAIStreamCancelsOnIdle(t *testing.T) {
-	withSSRFBypass(t)
-	ctx := t.Context()
-	// The server emits one valid chunk and then stalls. Without the
-	// watchdog, scanner.Scan() would hang forever. With the watchdog
-	// at 200ms, it must return a clear "stream idle" error in well
-	// under a second.
-	withLocalAIIdleTimeout(t, 200*time.Millisecond)
-
-	// hold blocks the handler until the test closes it. Register
-	// close(hold) FIRST so it runs LAST (defers are LIFO) — wait,
-	// that's the opposite. We want close(hold) to run BEFORE
-	// srv.Close() so the handler can return. Use t.Cleanup, which
-	// runs in reverse-registration order: register srv.Close first
-	// so it runs last, then close(hold) so it runs first.
-	hold := make(chan struct{})
-
-	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
-		w.Header().Set("Content-Type", "text/event-stream")
-		w.WriteHeader(http.StatusOK)
-		if f, ok := w.(http.Flusher); ok {
-			_, _ = io.WriteString(w, `data: {"choices":[{"delta":{"content":"hi"}}]}`+"\n")
-			f.Flush()
-		}
-		// Hang until either the client disconnects (watchdog cancels
-		// the request, which causes r.Context() to fire) or the test
-		// teardown signals via `hold`.
-		select {
-		case <-hold:
-		case <-r.Context().Done():
-		}
-	}))
-	t.Cleanup(srv.Close)
-	t.Cleanup(func() { close(hold) })
-
-	l := newLocalAIForTest(srv.URL)
-	var got []string
-	var mu sync.Mutex
-	err := l.ChatStreamlyWithSender(ctx, "gpt-4",
-		[]Message{{Role: "user", Content: "x"}},
-		&APIConfig{}, nil, nil,
-		func(content *string, _ *string) error {
-			if content == nil || *content == "" {
-				return nil
-			}
-			mu.Lock()
-			got = append(got, *content)
-			mu.Unlock()
-			return nil
-		},
-	)
-
-	if err == nil {
-		t.Fatal("expected an idle-timeout error, got nil")
-	}
-	if !strings.Contains(err.Error(), "idle for more than") {
-		t.Errorf("expected idle-timeout error, got %v", err)
-	}
-	mu.Lock()
-	defer mu.Unlock()
-	if len(got) == 0 || got[0] != "hi" {
-		t.Errorf("expected first chunk before stall, got %v", got)
-	}
-}
-
-func TestLocalAIStreamCompletesWithoutTriggeringWatchdog(t *testing.T) {
-	withSSRFBypass(t)
-	ctx := t.Context()
-	// Sanity check: a fast, complete stream should not trip the
-	// watchdog even with a moderately tight idle window.
-	withLocalAIIdleTimeout(t, 500*time.Millisecond)
-
-	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
-		w.Header().Set("Content-Type", "text/event-stream")
-		w.WriteHeader(http.StatusOK)
-		f, _ := w.(http.Flusher)
-		_, _ = io.WriteString(w,
-			`data: {"choices":[{"delta":{"content":"a"}}]}`+"\n"+
-				`data: {"choices":[{"delta":{"content":"b"}}]}`+"\n"+
-				`data: {"choices":[{"delta":{},"finish_reason":"stop"}]}`+"\n"+
-				`data: [DONE]`+"\n",
-		)
-		if f != nil {
-			f.Flush()
-		}
-	}))
-	defer srv.Close()
-
-	l := newLocalAIForTest(srv.URL)
-	var chunks []string
-	err := l.ChatStreamlyWithSender(ctx, "gpt-4",
-		[]Message{{Role: "user", Content: "x"}},
-		&APIConfig{}, nil, nil,
-		func(content *string, _ *string) error {
-			if content != nil && *content != "" && *content != "[DONE]" {
-				chunks = append(chunks, *content)
-			}
-			return nil
-		},
-	)
-	if err != nil {
-		t.Fatalf("stream: %v", err)
-	}
-	if strings.Join(chunks, "") != "ab" {
-		t.Errorf("chunks=%v want [a b]", chunks)
-	}
-}
-
 func TestLocalAIStreamRequiresSender(t *testing.T) {
 	withSSRFBypass(t)
 	ctx := t.Context()
@@ -335,54 +213,7 @@ func TestLocalAIEmbedEmptyInputShortCircuits(t *testing.T) {
 	}
 }
 
-// ---------- reasoning extraction (multi-field) ----------
-
-// Table-driven unit coverage of the helper. Mirrors the priority order
-// reasoning_content > reasoning > thinking declared in
-// localAIReasoningFields. New upstream field names can be added by
-// extending that slice without touching call sites.
-func TestExtractLocalAIReasoning(t *testing.T) {
-	cases := []struct {
-		name string
-		in   map[string]interface{}
-		want string
-	}{
-		{"empty map", map[string]interface{}{}, ""},
-		{"reasoning_content wins", map[string]interface{}{
-			"reasoning_content": "rc",
-			"reasoning":         "r",
-			"thinking":          "t",
-		}, "rc"},
-		{"reasoning when no reasoning_content", map[string]interface{}{
-			"reasoning": "r",
-			"thinking":  "t",
-		}, "r"},
-		{"thinking when only that is set", map[string]interface{}{
-			"thinking": "qwen3-thought",
-		}, "qwen3-thought"},
-		{"empty string treated as absent", map[string]interface{}{
-			"reasoning_content": "",
-			"reasoning":         "fallback",
-		}, "fallback"},
-		{"non-string ignored", map[string]interface{}{
-			"reasoning_content": 42,
-			"reasoning":         "fallback",
-		}, "fallback"},
-	}
-	for _, tc := range cases {
-		t.Run(tc.name, func(t *testing.T) {
-			got := extractLocalAIReasoning(tc.in)
-			if got != tc.want {
-				t.Errorf("got=%q want=%q", got, tc.want)
-			}
-		})
-	}
-}
-
-// Non-streaming chat against an upstream that puts the trace in
-// message.reasoning_content (kimi-k2.6, OpenAI o-series, DeepSeek-R1
-// when proxied through OpenAI-shim). The driver must surface it on
-// ChatResponse.ReasonContent.
+// ---------- reasoning content (message.reasoning_content) ----------
 func TestLocalAIChatExtractsReasoningContent(t *testing.T) {
 	withSSRFBypass(t)
 	ctx := t.Context()
@@ -411,33 +242,6 @@ func TestLocalAIChatExtractsReasoningContent(t *testing.T) {
 	}
 }
 
-// Non-streaming chat that uses message.thinking (Qwen3 via Ollama-shim
-// inside LocalAI). The driver must surface it on ReasonContent too.
-func TestLocalAIChatExtractsThinking(t *testing.T) {
-	withSSRFBypass(t)
-	ctx := t.Context()
-	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
-		_, _ = io.WriteString(w, `{"choices":[{"message":{
-			"role":"assistant",
-			"content":"12",
-			"thinking":"Compute 15/100 * 80"
-		}}]}`)
-	}))
-	defer srv.Close()
-
-	l := newLocalAIForTest(srv.URL)
-	resp, err := l.ChatWithMessages(ctx, "qwen3-32b",
-		[]Message{{Role: "user", Content: "15% of 80?"}},
-		&APIConfig{}, nil, nil,
-	)
-	if err != nil {
-		t.Fatalf("Chat: %v", err)
-	}
-	if *resp.ReasonContent != "Compute 15/100 * 80" {
-		t.Errorf("ReasonContent=%q want %q", *resp.ReasonContent, "Compute 15/100 * 80")
-	}
-}
-
 // Regression net: a response with no reasoning field at all (any
 // non-reasoning model) must produce empty ReasonContent without
 // crashing or erroring.
@@ -515,46 +319,6 @@ func TestLocalAIStreamExtractsReasoningContentDelta(t *testing.T) {
 	}
 }
 
-// Streaming chat where the upstream uses delta.thinking (Qwen3 shape).
-// The same handler must work.
-func TestLocalAIStreamExtractsThinkingDelta(t *testing.T) {
-	withSSRFBypass(t)
-	ctx := t.Context()
-	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
-		w.Header().Set("Content-Type", "text/event-stream")
-		w.WriteHeader(http.StatusOK)
-		_, _ = io.WriteString(w,
-			`data: {"choices":[{"index":0,"delta":{"thinking":"qwen-trace"}}]}`+"\n"+
-				`data: {"choices":[{"index":0,"delta":{"content":"final"},"finish_reason":"stop"}]}`+"\n"+
-				`data: [DONE]`+"\n",
-		)
-	}))
-	defer srv.Close()
-
-	l := newLocalAIForTest(srv.URL)
-	var got []string
-	err := l.ChatStreamlyWithSender(ctx, "qwen3-32b",
-		[]Message{{Role: "user", Content: "x"}},
-		&APIConfig{}, nil, nil,
-		func(c *string, r *string) error {
-			if r != nil && *r != "" {
-				got = append(got, "R:"+*r)
-			}
-			if c != nil && *c != "" && *c != "[DONE]" {
-				got = append(got, "C:"+*c)
-			}
-			return nil
-		},
-	)
-	if err != nil {
-		t.Fatalf("stream: %v", err)
-	}
-	want := []string{"R:qwen-trace", "C:final"}
-	if len(got) != 2 || got[0] != want[0] || got[1] != want[1] {
-		t.Errorf("seq=%v want %v", got, want)
-	}
-}
-
 // Request-side: ChatConfig.Effort must flow into request body as
 // reasoning_effort.
 func TestLocalAIChatPropagatesReasoningEffort(t *testing.T) {