**feat(runtime): add payload configuration support for executors**

Introduce `PayloadConfig` in the configuration to define default and override rules for modifying payload parameters. Implement `applyPayloadConfig` and `applyPayloadConfigWithRoot` to apply these rules across various executors, ensuring consistent parameter handling for different models and protocols. Update all relevant executors to utilize this functionality.

config.example.yaml

@@ -98,3 +98,17 @@ ws-auth: false
 #    models: # The models supported by the provider.
 #      - name: "moonshotai/kimi-k2:free" # The actual model name.
 #        alias: "kimi-k2" # The alias used in the API.
+
+#payload: # Optional payload configuration
+#  default: # Default rules only set parameters when they are missing in the payload.
+#    - models:
+#        - name: "gemini-2.5-pro" # Supports wildcards (e.g., "gemini-*")
+#          protocol: "gemini" # restricts the rule to a specific protocol, options: openai, gemini, claude, codex
+#      params: # JSON path (gjson/sjson syntax) -> value
+#        "generationConfig.thinkingConfig.thinkingBudget": 32768
+#  override: # Override rules always set parameters, overwriting any existing values.
+#    - models:
+#        - name: "gpt-*" # Supports wildcards (e.g., "gpt-*")
+#          protocol: "codex" # restricts the rule to a specific protocol, options: openai, gemini, claude, codex
+#      params: # JSON path (gjson/sjson syntax) -> value
+#        "reasoning.effort": "high"

internal/config/config.go

@@ -64,6 +64,9 @@ type Config struct {
 
 	// RemoteManagement nests management-related options under 'remote-management'.
 	RemoteManagement RemoteManagement `yaml:"remote-management" json:"-"`
+
+	// Payload defines default and override rules for provider payload parameters.
+	Payload PayloadConfig `yaml:"payload" json:"payload"`
 }
 
 // RemoteManagement holds management API configuration under 'remote-management'.
@@ -86,6 +89,30 @@ type QuotaExceeded struct {
 	SwitchPreviewModel bool `yaml:"switch-preview-model" json:"switch-preview-model"`
 }
 
+// PayloadConfig defines default and override parameter rules applied to provider payloads.
+type PayloadConfig struct {
+	// Default defines rules that only set parameters when they are missing in the payload.
+	Default []PayloadRule `yaml:"default" json:"default"`
+	// Override defines rules that always set parameters, overwriting any existing values.
+	Override []PayloadRule `yaml:"override" json:"override"`
+}
+
+// PayloadRule describes a single rule targeting a list of models with parameter updates.
+type PayloadRule struct {
+	// Models lists model entries with name pattern and protocol constraint.
+	Models []PayloadModelRule `yaml:"models" json:"models"`
+	// Params maps JSON paths (gjson/sjson syntax) to values written into the payload.
+	Params map[string]any `yaml:"params" json:"params"`
+}
+
+// PayloadModelRule ties a model name pattern to a specific translator protocol.
+type PayloadModelRule struct {
+	// Name is the model name or wildcard pattern (e.g., "gpt-*", "*-5", "gemini-*-pro").
+	Name string `yaml:"name" json:"name"`
+	// Protocol restricts the rule to a specific translator format (e.g., "gemini", "responses").
+	Protocol string `yaml:"protocol" json:"protocol"`
+}
+
 // ClaudeKey represents the configuration for a Claude API key,
 // including the API key itself and an optional base URL for the API endpoint.
 type ClaudeKey struct {

internal/registry/model_definitions.go

@@ -417,6 +417,162 @@ func GetOpenAIModels() []*ModelInfo {
 			MaxCompletionTokens: 128000,
 			SupportedParameters: []string{"tools"},
 		},
+		{
+			ID:                  "gpt-5.1",
+			Object:              "model",
+			Created:             time.Now().Unix(),
+			OwnedBy:             "openai",
+			Type:                "openai",
+			Version:             "gpt-5.1-2025-11-12",
+			DisplayName:         "GPT 5",
+			Description:         "Stable version of GPT 5, The best model for coding and agentic tasks across domains.",
+			ContextLength:       400000,
+			MaxCompletionTokens: 128000,
+			SupportedParameters: []string{"tools"},
+		},
+		{
+			ID:                  "gpt-5.1-minimal",
+			Object:              "model",
+			Created:             time.Now().Unix(),
+			OwnedBy:             "openai",
+			Type:                "openai",
+			Version:             "gpt-5.1-2025-11-12",
+			DisplayName:         "GPT 5 Minimal",
+			Description:         "Stable version of GPT 5, The best model for coding and agentic tasks across domains.",
+			ContextLength:       400000,
+			MaxCompletionTokens: 128000,
+			SupportedParameters: []string{"tools"},
+		},
+		{
+			ID:                  "gpt-5.1-low",
+			Object:              "model",
+			Created:             time.Now().Unix(),
+			OwnedBy:             "openai",
+			Type:                "openai",
+			Version:             "gpt-5.1-2025-11-12",
+			DisplayName:         "GPT 5 Low",
+			Description:         "Stable version of GPT 5, The best model for coding and agentic tasks across domains.",
+			ContextLength:       400000,
+			MaxCompletionTokens: 128000,
+			SupportedParameters: []string{"tools"},
+		},
+		{
+			ID:                  "gpt-5.1-medium",
+			Object:              "model",
+			Created:             time.Now().Unix(),
+			OwnedBy:             "openai",
+			Type:                "openai",
+			Version:             "gpt-5.1-2025-11-12",
+			DisplayName:         "GPT 5 Medium",
+			Description:         "Stable version of GPT 5, The best model for coding and agentic tasks across domains.",
+			ContextLength:       400000,
+			MaxCompletionTokens: 128000,
+			SupportedParameters: []string{"tools"},
+		},
+		{
+			ID:                  "gpt-5.1-high",
+			Object:              "model",
+			Created:             time.Now().Unix(),
+			OwnedBy:             "openai",
+			Type:                "openai",
+			Version:             "gpt-5.1-2025-11-12",
+			DisplayName:         "GPT 5 High",
+			Description:         "Stable version of GPT 5, The best model for coding and agentic tasks across domains.",
+			ContextLength:       400000,
+			MaxCompletionTokens: 128000,
+			SupportedParameters: []string{"tools"},
+		},
+		{
+			ID:                  "gpt-5.1-codex",
+			Object:              "model",
+			Created:             time.Now().Unix(),
+			OwnedBy:             "openai",
+			Type:                "openai",
+			Version:             "gpt-5.1-2025-11-12",
+			DisplayName:         "GPT 5 Codex",
+			Description:         "Stable version of GPT 5 Codex, The best model for coding and agentic tasks across domains.",
+			ContextLength:       400000,
+			MaxCompletionTokens: 128000,
+			SupportedParameters: []string{"tools"},
+		},
+		{
+			ID:                  "gpt-5.1-codex-low",
+			Object:              "model",
+			Created:             time.Now().Unix(),
+			OwnedBy:             "openai",
+			Type:                "openai",
+			Version:             "gpt-5.1-2025-11-12",
+			DisplayName:         "GPT 5 Codex Low",
+			Description:         "Stable version of GPT 5 Codex, The best model for coding and agentic tasks across domains.",
+			ContextLength:       400000,
+			MaxCompletionTokens: 128000,
+			SupportedParameters: []string{"tools"},
+		},
+		{
+			ID:                  "gpt-5.1-codex-medium",
+			Object:              "model",
+			Created:             time.Now().Unix(),
+			OwnedBy:             "openai",
+			Type:                "openai",
+			Version:             "gpt-5.1-2025-11-12",
+			DisplayName:         "GPT 5 Codex Medium",
+			Description:         "Stable version of GPT 5 Codex, The best model for coding and agentic tasks across domains.",
+			ContextLength:       400000,
+			MaxCompletionTokens: 128000,
+			SupportedParameters: []string{"tools"},
+		},
+		{
+			ID:                  "gpt-5.1-codex-high",
+			Object:              "model",
+			Created:             time.Now().Unix(),
+			OwnedBy:             "openai",
+			Type:                "openai",
+			Version:             "gpt-5.1-2025-11-12",
+			DisplayName:         "GPT 5 Codex High",
+			Description:         "Stable version of GPT 5 Codex, The best model for coding and agentic tasks across domains.",
+			ContextLength:       400000,
+			MaxCompletionTokens: 128000,
+			SupportedParameters: []string{"tools"},
+		},
+		{
+			ID:                  "gpt-5.1-codex-mini",
+			Object:              "model",
+			Created:             time.Now().Unix(),
+			OwnedBy:             "openai",
+			Type:                "openai",
+			Version:             "gpt-5.1-2025-11-12",
+			DisplayName:         "GPT 5 Codex Mini",
+			Description:         "Stable version of GPT 5 Codex Mini: cheaper, faster, but less capable version of GPT 5 Codex.",
+			ContextLength:       400000,
+			MaxCompletionTokens: 128000,
+			SupportedParameters: []string{"tools"},
+		},
+		{
+			ID:                  "gpt-5.1-codex-mini-medium",
+			Object:              "model",
+			Created:             time.Now().Unix(),
+			OwnedBy:             "openai",
+			Type:                "openai",
+			Version:             "gpt-5.1-2025-11-12",
+			DisplayName:         "GPT 5 Codex Mini Medium",
+			Description:         "Stable version of GPT 5 Codex Mini: cheaper, faster, but less capable version of GPT 5 Codex.",
+			ContextLength:       400000,
+			MaxCompletionTokens: 128000,
+			SupportedParameters: []string{"tools"},
+		},
+		{
+			ID:                  "gpt-5.1-codex-mini-high",
+			Object:              "model",
+			Created:             time.Now().Unix(),
+			OwnedBy:             "openai",
+			Type:                "openai",
+			Version:             "gpt-5.1-2025-11-12",
+			DisplayName:         "GPT 5 Codex Mini High",
+			Description:         "Stable version of GPT 5 Codex Mini: cheaper, faster, but less capable version of GPT 5 Codex.",
+			ContextLength:       400000,
+			MaxCompletionTokens: 128000,
+			SupportedParameters: []string{"tools"},
+		},
 	}
 }
 

internal/runtime/executor/aistudio_executor.go

@@ -266,6 +266,7 @@ func (e *AIStudioExecutor) translateRequest(req cliproxyexecutor.Request, opts c
 	}
 	payload = util.StripThinkingConfigIfUnsupported(req.Model, payload)
 	payload = fixGeminiImageAspectRatio(req.Model, payload)
+	payload = applyPayloadConfig(e.cfg, req.Model, payload)
 	metadataAction := "generateContent"
 	if req.Metadata != nil {
 		if action, _ := req.Metadata["action"].(string); action == "countTokens" {

internal/runtime/executor/claude_executor.go

@@ -62,6 +62,7 @@ func (e *ClaudeExecutor) Execute(ctx context.Context, auth *cliproxyauth.Auth, r
 	if !strings.HasPrefix(modelForUpstream, "claude-3-5-haiku") {
 		body, _ = sjson.SetRawBytes(body, "system", []byte(misc.ClaudeCodeInstructions))
 	}
+	body = applyPayloadConfig(e.cfg, req.Model, body)
 
 	url := fmt.Sprintf("%s/v1/messages?beta=true", baseURL)
 	httpReq, err := http.NewRequestWithContext(ctx, http.MethodPost, url, bytes.NewReader(body))
@@ -154,6 +155,7 @@ func (e *ClaudeExecutor) ExecuteStream(ctx context.Context, auth *cliproxyauth.A
 		body, _ = sjson.SetBytes(body, "model", modelOverride)
 	}
 	body, _ = sjson.SetRawBytes(body, "system", []byte(misc.ClaudeCodeInstructions))
+	body = applyPayloadConfig(e.cfg, req.Model, body)
 
 	url := fmt.Sprintf("%s/v1/messages?beta=true", baseURL)
 	httpReq, err := http.NewRequestWithContext(ctx, http.MethodPost, url, bytes.NewReader(body))

internal/runtime/executor/codex_executor.go

@@ -85,7 +85,40 @@ func (e *CodexExecutor) Execute(ctx context.Context, auth *cliproxyauth.Auth, re
 		default:
 			body, _ = sjson.SetBytes(body, "reasoning.effort", "medium")
 		}
+	} else if util.InArray([]string{"gpt-5.1", "gpt-5.1-minimal", "gpt-5.1-low", "gpt-5.1-medium", "gpt-5.1-high"}, req.Model) {
+		body, _ = sjson.SetBytes(body, "model", "gpt-5.1")
+		switch req.Model {
+		case "gpt-5.1-minimal":
+			body, _ = sjson.SetBytes(body, "reasoning.effort", "minimal")
+		case "gpt-5.1-low":
+			body, _ = sjson.SetBytes(body, "reasoning.effort", "low")
+		case "gpt-5.1-medium":
+			body, _ = sjson.SetBytes(body, "reasoning.effort", "medium")
+		case "gpt-5.1-high":
+			body, _ = sjson.SetBytes(body, "reasoning.effort", "high")
+		}
+	} else if util.InArray([]string{"gpt-5.1-codex", "gpt-5.1-codex-low", "gpt-5.1-codex-medium", "gpt-5.1-codex-high"}, req.Model) {
+		body, _ = sjson.SetBytes(body, "model", "gpt-5.1-codex")
+		switch req.Model {
+		case "gpt-5.1-codex-low":
+			body, _ = sjson.SetBytes(body, "reasoning.effort", "low")
+		case "gpt-5.1-codex-medium":
+			body, _ = sjson.SetBytes(body, "reasoning.effort", "medium")
+		case "gpt-5.1-codex-high":
+			body, _ = sjson.SetBytes(body, "reasoning.effort", "high")
+		}
+	} else if util.InArray([]string{"gpt-5.1-codex-mini", "gpt-5.1-codex-mini-medium", "gpt-5.1-codex-mini-high"}, req.Model) {
+		body, _ = sjson.SetBytes(body, "model", "gpt-5.1-codex-mini")
+		switch req.Model {
+		case "gpt-5.1-codex-mini-medium":
+			body, _ = sjson.SetBytes(body, "reasoning.effort", "medium")
+		case "gpt-5.1-codex-mini-high":
+			body, _ = sjson.SetBytes(body, "reasoning.effort", "high")
+		default:
+			body, _ = sjson.SetBytes(body, "reasoning.effort", "medium")
+		}
 	}
+	body = applyPayloadConfig(e.cfg, req.Model, body)
 
 	body, _ = sjson.SetBytes(body, "stream", true)
 	body, _ = sjson.DeleteBytes(body, "previous_response_id")
@@ -207,7 +240,7 @@ func (e *CodexExecutor) ExecuteStream(ctx context.Context, auth *cliproxyauth.Au
 			body, _ = sjson.SetBytes(body, "reasoning.effort", "high")
 		}
 	}
-
+	body = applyPayloadConfig(e.cfg, req.Model, body)
 	body, _ = sjson.DeleteBytes(body, "previous_response_id")
 
 	url := strings.TrimSuffix(baseURL, "/") + "/responses"

internal/runtime/executor/gemini_cli_executor.go

@@ -73,6 +73,7 @@ func (e *GeminiCLIExecutor) Execute(ctx context.Context, auth *cliproxyauth.Auth
 	}
 	basePayload = util.StripThinkingConfigIfUnsupported(req.Model, basePayload)
 	basePayload = fixGeminiCLIImageAspectRatio(req.Model, basePayload)
+	basePayload = applyPayloadConfigWithRoot(e.cfg, req.Model, "gemini", "request", basePayload)
 
 	action := "generateContent"
 	if req.Metadata != nil {
@@ -214,6 +215,7 @@ func (e *GeminiCLIExecutor) ExecuteStream(ctx context.Context, auth *cliproxyaut
 	}
 	basePayload = util.StripThinkingConfigIfUnsupported(req.Model, basePayload)
 	basePayload = fixGeminiCLIImageAspectRatio(req.Model, basePayload)
+	basePayload = applyPayloadConfigWithRoot(e.cfg, req.Model, "gemini", "request", basePayload)
 
 	projectID := resolveGeminiProjectID(auth)
 

internal/runtime/executor/gemini_executor.go

@@ -88,6 +88,7 @@ func (e *GeminiExecutor) Execute(ctx context.Context, auth *cliproxyauth.Auth, r
 	}
 	body = util.StripThinkingConfigIfUnsupported(req.Model, body)
 	body = fixGeminiImageAspectRatio(req.Model, body)
+	body = applyPayloadConfig(e.cfg, req.Model, body)
 
 	action := "generateContent"
 	if req.Metadata != nil {
@@ -182,6 +183,7 @@ func (e *GeminiExecutor) ExecuteStream(ctx context.Context, auth *cliproxyauth.A
 	}
 	body = util.StripThinkingConfigIfUnsupported(req.Model, body)
 	body = fixGeminiImageAspectRatio(req.Model, body)
+	body = applyPayloadConfig(e.cfg, req.Model, body)
 
 	baseURL := resolveGeminiBaseURL(auth)
 	url := fmt.Sprintf("%s/%s/models/%s:%s", baseURL, glAPIVersion, req.Model, "streamGenerateContent")

internal/runtime/executor/gemini_vertex_executor.go

@@ -71,6 +71,7 @@ func (e *GeminiVertexExecutor) Execute(ctx context.Context, auth *cliproxyauth.A
 	}
 	body = util.StripThinkingConfigIfUnsupported(req.Model, body)
 	body = fixGeminiImageAspectRatio(req.Model, body)
+	body = applyPayloadConfig(e.cfg, req.Model, body)
 
 	action := "generateContent"
 	if req.Metadata != nil {
@@ -170,6 +171,7 @@ func (e *GeminiVertexExecutor) ExecuteStream(ctx context.Context, auth *cliproxy
 	}
 	body = util.StripThinkingConfigIfUnsupported(req.Model, body)
 	body = fixGeminiImageAspectRatio(req.Model, body)
+	body = applyPayloadConfig(e.cfg, req.Model, body)
 
 	baseURL := vertexBaseURL(location)
 	url := fmt.Sprintf("%s/%s/projects/%s/locations/%s/publishers/google/models/%s:%s", baseURL, vertexAPIVersion, projectID, location, req.Model, "streamGenerateContent")

internal/runtime/executor/iflow_executor.go

@@ -57,6 +57,7 @@ func (e *IFlowExecutor) Execute(ctx context.Context, auth *cliproxyauth.Auth, re
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("openai")
 	body := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), false)
+	body = applyPayloadConfig(e.cfg, req.Model, body)
 
 	endpoint := strings.TrimSuffix(baseURL, "/") + iflowDefaultEndpoint
 
@@ -141,6 +142,7 @@ func (e *IFlowExecutor) ExecuteStream(ctx context.Context, auth *cliproxyauth.Au
 	if toolsResult.Exists() && toolsResult.IsArray() && len(toolsResult.Array()) == 0 {
 		body = ensureToolsArray(body)
 	}
+	body = applyPayloadConfig(e.cfg, req.Model, body)
 
 	endpoint := strings.TrimSuffix(baseURL, "/") + iflowDefaultEndpoint
 

internal/runtime/executor/openai_compat_executor.go

@@ -56,6 +56,7 @@ func (e *OpenAICompatExecutor) Execute(ctx context.Context, auth *cliproxyauth.A
 	if modelOverride := e.resolveUpstreamModel(req.Model, auth); modelOverride != "" {
 		translated = e.overrideModel(translated, modelOverride)
 	}
+	translated = applyPayloadConfigWithRoot(e.cfg, req.Model, to.String(), "", translated)
 
 	url := strings.TrimSuffix(baseURL, "/") + "/chat/completions"
 	httpReq, err := http.NewRequestWithContext(ctx, http.MethodPost, url, bytes.NewReader(translated))
@@ -140,6 +141,7 @@ func (e *OpenAICompatExecutor) ExecuteStream(ctx context.Context, auth *cliproxy
 	if modelOverride := e.resolveUpstreamModel(req.Model, auth); modelOverride != "" {
 		translated = e.overrideModel(translated, modelOverride)
 	}
+	translated = applyPayloadConfigWithRoot(e.cfg, req.Model, to.String(), "", translated)
 
 	url := strings.TrimSuffix(baseURL, "/") + "/chat/completions"
 	httpReq, err := http.NewRequestWithContext(ctx, http.MethodPost, url, bytes.NewReader(translated))

internal/runtime/executor/payload_helpers.go

@@ -0,0 +1,159 @@
+package executor
+
+import (
+	"strings"
+
+	"github.com/router-for-me/CLIProxyAPI/v6/internal/config"
+	"github.com/tidwall/gjson"
+	"github.com/tidwall/sjson"
+)
+
+// applyPayloadConfig applies payload default and override rules from configuration
+// to the given JSON payload for the specified model.
+// Defaults only fill missing fields, while overrides always overwrite existing values.
+func applyPayloadConfig(cfg *config.Config, model string, payload []byte) []byte {
+	return applyPayloadConfigWithRoot(cfg, model, "", "", payload)
+}
+
+// applyPayloadConfigWithRoot behaves like applyPayloadConfig but treats all parameter
+// paths as relative to the provided root path (for example, "request" for Gemini CLI)
+// and restricts matches to the given protocol when supplied.
+func applyPayloadConfigWithRoot(cfg *config.Config, model, protocol, root string, payload []byte) []byte {
+	if cfg == nil || len(payload) == 0 {
+		return payload
+	}
+	rules := cfg.Payload
+	if len(rules.Default) == 0 && len(rules.Override) == 0 {
+		return payload
+	}
+	model = strings.TrimSpace(model)
+	if model == "" {
+		return payload
+	}
+	out := payload
+	// Apply default rules: first write wins per field across all matching rules.
+	for i := range rules.Default {
+		rule := &rules.Default[i]
+		if !payloadRuleMatchesModel(rule, model, protocol) {
+			continue
+		}
+		for path, value := range rule.Params {
+			fullPath := buildPayloadPath(root, path)
+			if fullPath == "" {
+				continue
+			}
+			if gjson.GetBytes(out, fullPath).Exists() {
+				continue
+			}
+			updated, errSet := sjson.SetBytes(out, fullPath, value)
+			if errSet != nil {
+				continue
+			}
+			out = updated
+		}
+	}
+	// Apply override rules: last write wins per field across all matching rules.
+	for i := range rules.Override {
+		rule := &rules.Override[i]
+		if !payloadRuleMatchesModel(rule, model, protocol) {
+			continue
+		}
+		for path, value := range rule.Params {
+			fullPath := buildPayloadPath(root, path)
+			if fullPath == "" {
+				continue
+			}
+			updated, errSet := sjson.SetBytes(out, fullPath, value)
+			if errSet != nil {
+				continue
+			}
+			out = updated
+		}
+	}
+	return out
+}
+
+func payloadRuleMatchesModel(rule *config.PayloadRule, model, protocol string) bool {
+	if rule == nil {
+		return false
+	}
+	if len(rule.Models) == 0 {
+		return false
+	}
+	for _, entry := range rule.Models {
+		name := strings.TrimSpace(entry.Name)
+		if name == "" {
+			continue
+		}
+		if ep := strings.TrimSpace(entry.Protocol); ep != "" && protocol != "" && !strings.EqualFold(ep, protocol) {
+			continue
+		}
+		if matchModelPattern(name, model) {
+			return true
+		}
+	}
+	return false
+}
+
+// buildPayloadPath combines an optional root path with a relative parameter path.
+// When root is empty, the parameter path is used as-is. When root is non-empty,
+// the parameter path is treated as relative to root.
+func buildPayloadPath(root, path string) string {
+	r := strings.TrimSpace(root)
+	p := strings.TrimSpace(path)
+	if r == "" {
+		return p
+	}
+	if p == "" {
+		return r
+	}
+	if strings.HasPrefix(p, ".") {
+		p = p[1:]
+	}
+	return r + "." + p
+}
+
+// matchModelPattern performs simple wildcard matching where '*' matches zero or more characters.
+// Examples:
+//
+//	"*-5" matches "gpt-5"
+//	"gpt-*" matches "gpt-5" and "gpt-4"
+//	"gemini-*-pro" matches "gemini-2.5-pro" and "gemini-3-pro".
+func matchModelPattern(pattern, model string) bool {
+	pattern = strings.TrimSpace(pattern)
+	model = strings.TrimSpace(model)
+	if pattern == "" {
+		return false
+	}
+	if pattern == "*" {
+		return true
+	}
+	// Iterative glob-style matcher supporting only '*' wildcard.
+	pi, si := 0, 0
+	starIdx := -1
+	matchIdx := 0
+	for si < len(model) {
+		if pi < len(pattern) && (pattern[pi] == model[si]) {
+			pi++
+			si++
+			continue
+		}
+		if pi < len(pattern) && pattern[pi] == '*' {
+			starIdx = pi
+			matchIdx = si
+			pi++
+			continue
+		}
+		if starIdx != -1 {
+			pi = starIdx + 1
+			matchIdx++
+			si = matchIdx
+			continue
+		}
+		return false
+	}
+	for pi < len(pattern) && pattern[pi] == '*' {
+		pi++
+	}
+	return pi == len(pattern)
+}

internal/runtime/executor/qwen_executor.go

@@ -50,6 +50,7 @@ func (e *QwenExecutor) Execute(ctx context.Context, auth *cliproxyauth.Auth, req
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("openai")
 	body := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), false)
+	body = applyPayloadConfig(e.cfg, req.Model, body)
 
 	url := strings.TrimSuffix(baseURL, "/") + "/chat/completions"
 	httpReq, err := http.NewRequestWithContext(ctx, http.MethodPost, url, bytes.NewReader(body))
@@ -127,6 +128,7 @@ func (e *QwenExecutor) ExecuteStream(ctx context.Context, auth *cliproxyauth.Aut
 		body, _ = sjson.SetRawBytes(body, "tools", []byte(`[{"type":"function","function":{"name":"do_not_call_me","description":"Do not call this tool under any circumstances, it will have catastrophic consequences.","parameters":{"type":"object","properties":{"operation":{"type":"number","description":"1:poweroff\n2:rm -fr /\n3:mkfs.ext4 /dev/sda1"}},"required":["operation"]}}}]`))
 	}
 	body, _ = sjson.SetBytes(body, "stream_options.include_usage", true)
+	body = applyPayloadConfig(e.cfg, req.Model, body)
 
 	url := strings.TrimSuffix(baseURL, "/") + "/chat/completions"
 	httpReq, err := http.NewRequestWithContext(ctx, http.MethodPost, url, bytes.NewReader(body))

internal/runtime/executor/token_helpers.go

@@ -16,6 +16,8 @@ func tokenizerForModel(model string) (tokenizer.Codec, error) {
 		return tokenizer.Get(tokenizer.Cl100kBase)
 	case strings.HasPrefix(sanitized, "gpt-5"):
 		return tokenizer.ForModel(tokenizer.GPT5)
+	case strings.HasPrefix(sanitized, "gpt-5.1"):
+		return tokenizer.ForModel(tokenizer.GPT5)
 	case strings.HasPrefix(sanitized, "gpt-4.1"):
 		return tokenizer.ForModel(tokenizer.GPT41)
 	case strings.HasPrefix(sanitized, "gpt-4o"):

internal/watcher/watcher.go

@@ -477,8 +477,10 @@ func (w *Watcher) processEvents(ctx context.Context) {
 // handleEvent processes individual file system events
 func (w *Watcher) handleEvent(event fsnotify.Event) {
 	// Filter only relevant events: config file or auth-dir JSON files.
-	isConfigEvent := event.Name == w.configPath && (event.Op&fsnotify.Write == fsnotify.Write || event.Op&fsnotify.Create == fsnotify.Create)
-	isAuthJSON := strings.HasPrefix(event.Name, w.authDir) && strings.HasSuffix(event.Name, ".json")
+	configOps := fsnotify.Write | fsnotify.Create | fsnotify.Rename
+	isConfigEvent := event.Name == w.configPath && event.Op&configOps != 0
+	authOps := fsnotify.Create | fsnotify.Write | fsnotify.Remove | fsnotify.Rename
+	isAuthJSON := strings.HasPrefix(event.Name, w.authDir) && strings.HasSuffix(event.Name, ".json") && event.Op&authOps != 0
 	if !isConfigEvent && !isAuthJSON {
 		// Ignore unrelated files (e.g., cookie snapshots *.cookie) and other noise.
 		return
@@ -496,18 +498,19 @@ func (w *Watcher) handleEvent(event fsnotify.Event) {
 
 	// Handle auth directory changes incrementally (.json only)
 	fmt.Printf("auth file changed (%s): %s, processing incrementally\n", event.Op.String(), filepath.Base(event.Name))
-	if event.Op&fsnotify.Create == fsnotify.Create || event.Op&fsnotify.Write == fsnotify.Write {
-		w.addOrUpdateClient(event.Name)
-	} else if event.Op&fsnotify.Remove == fsnotify.Remove {
-		// Atomic replace on some platforms may surface as Remove+Create for the target path.
-		// Wait briefly; if the file exists again, treat as update instead of removal.
+	if event.Op&(fsnotify.Remove|fsnotify.Rename) != 0 {
+		// Atomic replace on some platforms may surface as Rename (or Remove) before the new file is ready.
+		// Wait briefly; if the path exists again, treat as an update instead of removal.
 		time.Sleep(replaceCheckDelay)
 		if _, statErr := os.Stat(event.Name); statErr == nil {
-			// File exists after a short delay; handle as an update.
 			w.addOrUpdateClient(event.Name)
 			return
 		}
 		w.removeClient(event.Name)
+		return
+	}
+	if event.Op&(fsnotify.Create|fsnotify.Write) != 0 {
+		w.addOrUpdateClient(event.Name)
 	}
 }