feat(registry, executor): add support for glm-4.6 model and enhance Gemini CLI token handling

- Added `glm-4.6` model to registry and documentation.
- Updated Gemini CLI executor to pass configuration to `prepareGeminiCLITokenSource` for improved token management.

README.md

@@ -258,6 +258,7 @@ console.log(await claudeResponse.json());
 - gemini-2.5-pro
 - gemini-2.5-flash
 - gemini-2.5-flash-lite
+- gemini-2.5-flash-image
 - gemini-2.5-flash-image-preview
 - gpt-5
 - gpt-5-codex
@@ -277,6 +278,7 @@ console.log(await claudeResponse.json());
 - deepseek-v3
 - kimi-k2
 - glm-4.5
+- glm-4.6
 - tstars2.0
 - And other iFlow-supported models
 - Gemini models auto-switch to preview variants when needed

README_CN.md

@@ -270,6 +270,7 @@ console.log(await claudeResponse.json());
 - gemini-2.5-pro
 - gemini-2.5-flash
 - gemini-2.5-flash-lite
+- gemini-2.5-flash-image
 - gemini-2.5-flash-image-preview
 - gpt-5
 - gpt-5-codex
@@ -289,6 +290,7 @@ console.log(await claudeResponse.json());
 - deepseek-v3
 - kimi-k2
 - glm-4.5
+- glm-4.6
 - tstars2.0
 - 以及其他 iFlow 支持的模型
 - Gemini 模型在需要时自动切换到对应的 preview 版本

internal/registry/model_definitions.go

@@ -166,6 +166,20 @@ func GetGeminiCLIModels() []*ModelInfo {
 			OutputTokenLimit:           8192,
 			SupportedGenerationMethods: []string{"generateContent", "countTokens", "createCachedContent", "batchGenerateContent"},
 		},
+		{
+			ID:                         "gemini-2.5-flash-image",
+			Object:                     "model",
+			Created:                    time.Now().Unix(),
+			OwnedBy:                    "google",
+			Type:                       "gemini",
+			Name:                       "models/gemini-2.5-flash-image",
+			Version:                    "2.5",
+			DisplayName:                "Gemini 2.5 Flash Image",
+			Description:                "State-of-the-art image generation and editing model.",
+			InputTokenLimit:            1048576,
+			OutputTokenLimit:           8192,
+			SupportedGenerationMethods: []string{"generateContent", "countTokens", "createCachedContent", "batchGenerateContent"},
+		},
 	}
 }
 
@@ -354,6 +368,7 @@ func GetIFlowModels() []*ModelInfo {
 		{ID: "qwen3-max-preview", DisplayName: "Qwen3-Max-Preview", Description: "Qwen3 Max preview build"},
 		{ID: "kimi-k2-0905", DisplayName: "Kimi-K2-Instruct-0905", Description: "Moonshot Kimi K2 instruct 0905"},
 		{ID: "glm-4.5", DisplayName: "GLM-4.5", Description: "Zhipu GLM 4.5 general model"},
+		{ID: "glm-4.6", DisplayName: "GLM-4.6", Description: "Zhipu GLM 4.6 general model"},
 		{ID: "kimi-k2", DisplayName: "Kimi-K2", Description: "Moonshot Kimi K2 general model"},
 		{ID: "deepseek-v3.2", DisplayName: "DeepSeek-V3.2-Exp", Description: "DeepSeek V3.2 experimental"},
 		{ID: "deepseek-v3.1", DisplayName: "DeepSeek-V3.1-Terminus", Description: "DeepSeek V3.1 Terminus"},

internal/runtime/executor/gemini_cli_executor.go

@@ -51,7 +51,7 @@ func (e *GeminiCLIExecutor) Identifier() string { return "gemini-cli" }
 func (e *GeminiCLIExecutor) PrepareRequest(_ *http.Request, _ *cliproxyauth.Auth) error { return nil }
 
 func (e *GeminiCLIExecutor) Execute(ctx context.Context, auth *cliproxyauth.Auth, req cliproxyexecutor.Request, opts cliproxyexecutor.Options) (cliproxyexecutor.Response, error) {
-	tokenSource, baseTokenData, err := prepareGeminiCLITokenSource(ctx, auth)
+	tokenSource, baseTokenData, err := prepareGeminiCLITokenSource(ctx, e.cfg, auth)
 	if err != nil {
 		return cliproxyexecutor.Response{}, err
 	}
@@ -139,7 +139,7 @@ func (e *GeminiCLIExecutor) Execute(ctx context.Context, auth *cliproxyauth.Auth
 }
 
 func (e *GeminiCLIExecutor) ExecuteStream(ctx context.Context, auth *cliproxyauth.Auth, req cliproxyexecutor.Request, opts cliproxyexecutor.Options) (<-chan cliproxyexecutor.StreamChunk, error) {
-	tokenSource, baseTokenData, err := prepareGeminiCLITokenSource(ctx, auth)
+	tokenSource, baseTokenData, err := prepareGeminiCLITokenSource(ctx, e.cfg, auth)
 	if err != nil {
 		return nil, err
 	}
@@ -270,7 +270,7 @@ func (e *GeminiCLIExecutor) ExecuteStream(ctx context.Context, auth *cliproxyaut
 }
 
 func (e *GeminiCLIExecutor) CountTokens(ctx context.Context, auth *cliproxyauth.Auth, req cliproxyexecutor.Request, opts cliproxyexecutor.Options) (cliproxyexecutor.Response, error) {
-	tokenSource, baseTokenData, err := prepareGeminiCLITokenSource(ctx, auth)
+	tokenSource, baseTokenData, err := prepareGeminiCLITokenSource(ctx, e.cfg, auth)
 	if err != nil {
 		return cliproxyexecutor.Response{}, err
 	}
@@ -351,7 +351,7 @@ func (e *GeminiCLIExecutor) Refresh(ctx context.Context, auth *cliproxyauth.Auth
 	return auth, nil
 }
 
-func prepareGeminiCLITokenSource(ctx context.Context, auth *cliproxyauth.Auth) (oauth2.TokenSource, map[string]any, error) {
+func prepareGeminiCLITokenSource(ctx context.Context, cfg *config.Config, auth *cliproxyauth.Auth) (oauth2.TokenSource, map[string]any, error) {
 	if auth == nil || auth.Metadata == nil {
 		return nil, nil, fmt.Errorf("gemini-cli auth metadata missing")
 	}
@@ -395,8 +395,8 @@ func prepareGeminiCLITokenSource(ctx context.Context, auth *cliproxyauth.Auth) (
 	}
 
 	ctxToken := ctx
-	if rt, ok := ctx.Value("cliproxy.roundtripper").(http.RoundTripper); ok && rt != nil {
-		ctxToken = context.WithValue(ctxToken, oauth2.HTTPClient, &http.Client{Transport: rt})
+	if httpClient := newProxyAwareHTTPClient(ctx, cfg, auth, 0); httpClient != nil {
+		ctxToken = context.WithValue(ctxToken, oauth2.HTTPClient, httpClient)
 	}
 
 	src := conf.TokenSource(ctxToken, &token)
@@ -518,7 +518,7 @@ func geminiModelDisallowsThinking(model string) bool {
 		return false
 	}
 	lower := strings.ToLower(model)
-	for _, marker := range []string{"gemini-2.5-flash-image-preview"} {
+	for _, marker := range []string{"gemini-2.5-flash-image-preview", "gemini-2.5-flash-image"} {
 		if strings.Contains(lower, marker) {
 			return true
 		}

internal/translator/gemini-cli/openai/chat-completions/cli_openai_request.go

@@ -26,6 +26,20 @@ import (
 //   - []byte: The transformed request data in Gemini CLI API format
 func ConvertOpenAIRequestToGeminiCLI(modelName string, inputRawJSON []byte, _ bool) []byte {
 	rawJSON := bytes.Clone(inputRawJSON)
+	var pathsToDelete []string
+	root := gjson.ParseBytes(rawJSON)
+	util.Walk(root, "", "additionalProperties", &pathsToDelete)
+	util.Walk(root, "", "$schema", &pathsToDelete)
+	util.Walk(root, "", "ref", &pathsToDelete)
+
+	var err error
+	for _, p := range pathsToDelete {
+		rawJSON, err = sjson.DeleteBytes(rawJSON, p)
+		if err != nil {
+			continue
+		}
+	}
+
 	// Base envelope
 	out := []byte(`{"project":"","request":{"contents":[],"generationConfig":{"thinkingConfig":{"include_thoughts":true}}},"model":"gemini-2.5-pro"}`)
 
@@ -65,6 +79,31 @@ func ConvertOpenAIRequestToGeminiCLI(modelName string, inputRawJSON []byte, _ bo
 		out, _ = sjson.SetBytes(out, "request.generationConfig.topK", tkr.Num)
 	}
 
+	// Map OpenAI modalities -> Gemini CLI request.generationConfig.responseModalities
+	// e.g. "modalities": ["image", "text"] -> ["Image", "Text"]
+	if mods := gjson.GetBytes(rawJSON, "modalities"); mods.Exists() && mods.IsArray() {
+		var responseMods []string
+		for _, m := range mods.Array() {
+			switch strings.ToLower(m.String()) {
+			case "text":
+				responseMods = append(responseMods, "Text")
+			case "image":
+				responseMods = append(responseMods, "Image")
+			}
+		}
+		if len(responseMods) > 0 {
+			out, _ = sjson.SetBytes(out, "request.generationConfig.responseModalities", responseMods)
+		}
+	}
+
+	// OpenRouter-style image_config support
+	// If the input uses top-level image_config.aspect_ratio, map it into request.generationConfig.imageConfig.aspectRatio.
+	if imgCfg := gjson.GetBytes(rawJSON, "image_config"); imgCfg.Exists() && imgCfg.IsObject() {
+		if ar := imgCfg.Get("aspect_ratio"); ar.Exists() && ar.Type == gjson.String {
+			out, _ = sjson.SetBytes(out, "request.generationConfig.imageConfig.aspectRatio", ar.Str)
+		}
+	}
+
 	// messages -> systemInstruction + contents
 	messages := gjson.GetBytes(rawJSON, "messages")
 	if messages.IsArray() {
@@ -232,7 +271,7 @@ func ConvertOpenAIRequestToGeminiCLI(modelName string, inputRawJSON []byte, _ bo
 	}
 
 	var pathsToType []string
-	root := gjson.ParseBytes(out)
+	root = gjson.ParseBytes(out)
 	util.Walk(root, "", "type", &pathsToType)
 	for _, p := range pathsToType {
 		typeResult := gjson.GetBytes(out, p)

internal/translator/gemini-cli/openai/chat-completions/cli_openai_response.go

@@ -20,6 +20,7 @@ import (
 // convertCliResponseToOpenAIChatParams holds parameters for response conversion.
 type convertCliResponseToOpenAIChatParams struct {
 	UnixTimestamp int64
+	FunctionIndex int
 }
 
 // ConvertCliResponseToOpenAI translates a single chunk of a streaming response from the
@@ -40,6 +41,7 @@ func ConvertCliResponseToOpenAI(_ context.Context, _ string, originalRequestRawJ
 	if *param == nil {
 		*param = &convertCliResponseToOpenAIChatParams{
 			UnixTimestamp: 0,
+			FunctionIndex: 0,
 		}
 	}
 
@@ -117,13 +119,18 @@ func ConvertCliResponseToOpenAI(_ context.Context, _ string, originalRequestRawJ
 			} else if functionCallResult.Exists() {
 				// Handle function call content.
 				toolCallsResult := gjson.Get(template, "choices.0.delta.tool_calls")
-				if !toolCallsResult.Exists() || !toolCallsResult.IsArray() {
+				functionCallIndex := (*param).(*convertCliResponseToOpenAIChatParams).FunctionIndex
+				(*param).(*convertCliResponseToOpenAIChatParams).FunctionIndex++
+				if toolCallsResult.Exists() && toolCallsResult.IsArray() {
+					functionCallIndex = len(toolCallsResult.Array())
+				} else {
 					template, _ = sjson.SetRaw(template, "choices.0.delta.tool_calls", `[]`)
 				}
 
-				functionCallTemplate := `{"id": "","type": "function","function": {"name": "","arguments": ""}}`
+				functionCallTemplate := `{"id": "","index": 0,"type": "function","function": {"name": "","arguments": ""}}`
 				fcName := functionCallResult.Get("name").String()
 				functionCallTemplate, _ = sjson.Set(functionCallTemplate, "id", fmt.Sprintf("%s-%d", fcName, time.Now().UnixNano()))
+				functionCallTemplate, _ = sjson.Set(functionCallTemplate, "index", functionCallIndex)
 				functionCallTemplate, _ = sjson.Set(functionCallTemplate, "function.name", fcName)
 				if fcArgsResult := functionCallResult.Get("args"); fcArgsResult.Exists() {
 					functionCallTemplate, _ = sjson.Set(functionCallTemplate, "function.arguments", fcArgsResult.Raw)

internal/translator/gemini/openai/chat-completions/gemini_openai_request.go

@@ -26,6 +26,20 @@ import (
 //   - []byte: The transformed request data in Gemini API format
 func ConvertOpenAIRequestToGemini(modelName string, inputRawJSON []byte, _ bool) []byte {
 	rawJSON := bytes.Clone(inputRawJSON)
+	var pathsToDelete []string
+	root := gjson.ParseBytes(rawJSON)
+	util.Walk(root, "", "additionalProperties", &pathsToDelete)
+	util.Walk(root, "", "$schema", &pathsToDelete)
+	util.Walk(root, "", "ref", &pathsToDelete)
+
+	var err error
+	for _, p := range pathsToDelete {
+		rawJSON, err = sjson.DeleteBytes(rawJSON, p)
+		if err != nil {
+			continue
+		}
+	}
+
 	// Base envelope
 	out := []byte(`{"contents":[],"generationConfig":{"thinkingConfig":{"include_thoughts":true}}}`)
 
@@ -65,6 +79,31 @@ func ConvertOpenAIRequestToGemini(modelName string, inputRawJSON []byte, _ bool)
 		out, _ = sjson.SetBytes(out, "generationConfig.topK", tkr.Num)
 	}
 
+	// Map OpenAI modalities -> Gemini generationConfig.responseModalities
+	// e.g. "modalities": ["image", "text"] -> ["Image", "Text"]
+	if mods := gjson.GetBytes(rawJSON, "modalities"); mods.Exists() && mods.IsArray() {
+		var responseMods []string
+		for _, m := range mods.Array() {
+			switch strings.ToLower(m.String()) {
+			case "text":
+				responseMods = append(responseMods, "Text")
+			case "image":
+				responseMods = append(responseMods, "Image")
+			}
+		}
+		if len(responseMods) > 0 {
+			out, _ = sjson.SetBytes(out, "generationConfig.responseModalities", responseMods)
+		}
+	}
+
+	// OpenRouter-style image_config support
+	// If the input uses top-level image_config.aspect_ratio, map it into generationConfig.imageConfig.aspectRatio.
+	if imgCfg := gjson.GetBytes(rawJSON, "image_config"); imgCfg.Exists() && imgCfg.IsObject() {
+		if ar := imgCfg.Get("aspect_ratio"); ar.Exists() && ar.Type == gjson.String {
+			out, _ = sjson.SetBytes(out, "generationConfig.imageConfig.aspectRatio", ar.Str)
+		}
+	}
+
 	// messages -> systemInstruction + contents
 	messages := gjson.GetBytes(rawJSON, "messages")
 	if messages.IsArray() {
@@ -257,7 +296,7 @@ func ConvertOpenAIRequestToGemini(modelName string, inputRawJSON []byte, _ bool)
 	}
 
 	var pathsToType []string
-	root := gjson.ParseBytes(out)
+	root = gjson.ParseBytes(out)
 	util.Walk(root, "", "type", &pathsToType)
 	for _, p := range pathsToType {
 		typeResult := gjson.GetBytes(out, p)

internal/translator/gemini/openai/chat-completions/gemini_openai_response.go

@@ -19,6 +19,7 @@ import (
 // convertGeminiResponseToOpenAIChatParams holds parameters for response conversion.
 type convertGeminiResponseToOpenAIChatParams struct {
 	UnixTimestamp int64
+	FunctionIndex int
 }
 
 // ConvertGeminiResponseToOpenAI translates a single chunk of a streaming response from the
@@ -39,6 +40,7 @@ func ConvertGeminiResponseToOpenAI(_ context.Context, _ string, originalRequestR
 	if *param == nil {
 		*param = &convertGeminiResponseToOpenAIChatParams{
 			UnixTimestamp: 0,
+			FunctionIndex: 0,
 		}
 	}
 
@@ -120,13 +122,18 @@ func ConvertGeminiResponseToOpenAI(_ context.Context, _ string, originalRequestR
 			} else if functionCallResult.Exists() {
 				// Handle function call content.
 				toolCallsResult := gjson.Get(template, "choices.0.delta.tool_calls")
-				if !toolCallsResult.Exists() || !toolCallsResult.IsArray() {
+				functionCallIndex := (*param).(*convertGeminiResponseToOpenAIChatParams).FunctionIndex
+				(*param).(*convertGeminiResponseToOpenAIChatParams).FunctionIndex++
+				if toolCallsResult.Exists() && toolCallsResult.IsArray() {
+					functionCallIndex = len(toolCallsResult.Array())
+				} else {
 					template, _ = sjson.SetRaw(template, "choices.0.delta.tool_calls", `[]`)
 				}
 
-				functionCallTemplate := `{"id": "","type": "function","function": {"name": "","arguments": ""}}`
+				functionCallTemplate := `{"id": "","index": 0,"type": "function","function": {"name": "","arguments": ""}}`
 				fcName := functionCallResult.Get("name").String()
 				functionCallTemplate, _ = sjson.Set(functionCallTemplate, "id", fmt.Sprintf("%s-%d", fcName, time.Now().UnixNano()))
+				functionCallTemplate, _ = sjson.Set(functionCallTemplate, "index", functionCallIndex)
 				functionCallTemplate, _ = sjson.Set(functionCallTemplate, "function.name", fcName)
 				if fcArgsResult := functionCallResult.Get("args"); fcArgsResult.Exists() {
 					functionCallTemplate, _ = sjson.Set(functionCallTemplate, "function.arguments", fcArgsResult.Raw)