feat: openai response /v1/response/compact (#2644)

* feat: openai response /v1/response/compact

* feat: /v1/response/compact bill

* feat: /v1/response/compact

* feat: /v1/responses/compact -> codex channel

* feat: /v1/responses/compact -> codex channel

* feat: /v1/responses/compact -> codex channel

* feat: codex channel default models

* feat: compact model price

* feat: /v1/responses/comapct test

relay/channel/codex/adaptor.go

@@ -53,6 +53,8 @@ func (a *Adaptor) ConvertEmbeddingRequest(c *gin.Context, info *relaycommon.Rela
 }
 
 func (a *Adaptor) ConvertOpenAIResponsesRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.OpenAIResponsesRequest) (any, error) {
+	isCompact := info != nil && info.RelayMode == relayconstant.RelayModeResponsesCompact
+
 	if info != nil && info.ChannelSetting.SystemPrompt != "" {
 		systemPrompt := info.ChannelSetting.SystemPrompt
 
@@ -88,7 +90,9 @@ func (a *Adaptor) ConvertOpenAIResponsesRequest(c *gin.Context, info *relaycommo
 			}
 		}
 	}
-
+	if isCompact {
+		return request, nil
+	}
 	// codex: store must be false
 	request.Store = json.RawMessage("false")
 	// rm max_output_tokens
@@ -102,10 +106,14 @@ func (a *Adaptor) DoRequest(c *gin.Context, info *relaycommon.RelayInfo, request
 }
 
 func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, info *relaycommon.RelayInfo) (usage any, err *types.NewAPIError) {
-	if info.RelayMode != relayconstant.RelayModeResponses {
+	if info.RelayMode != relayconstant.RelayModeResponses && info.RelayMode != relayconstant.RelayModeResponsesCompact {
 		return nil, types.NewError(errors.New("codex channel: endpoint not supported"), types.ErrorCodeInvalidRequest)
 	}
 
+	if info.RelayMode == relayconstant.RelayModeResponsesCompact {
+		return openai.OaiResponsesCompactionHandler(c, resp)
+	}
+
 	if info.IsStream {
 		return openai.OaiResponsesStreamHandler(c, info, resp)
 	}
@@ -121,10 +129,14 @@ func (a *Adaptor) GetChannelName() string {
 }
 
 func (a *Adaptor) GetRequestURL(info *relaycommon.RelayInfo) (string, error) {
-	if info.RelayMode != relayconstant.RelayModeResponses {
-		return "", errors.New("codex channel: only /v1/responses is supported")
+	if info.RelayMode != relayconstant.RelayModeResponses && info.RelayMode != relayconstant.RelayModeResponsesCompact {
+		return "", errors.New("codex channel: only /v1/responses and /v1/responses/compact are supported")
 	}
-	return relaycommon.GetFullRequestURL(info.ChannelBaseUrl, "/backend-api/codex/responses", info.ChannelType), nil
+	path := "/backend-api/codex/responses"
+	if info.RelayMode == relayconstant.RelayModeResponsesCompact {
+		path = "/backend-api/codex/responses/compact"
+	}
+	return relaycommon.GetFullRequestURL(info.ChannelBaseUrl, path, info.ChannelType), nil
 }
 
 func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Header, info *relaycommon.RelayInfo) error {

relay/channel/codex/constants.go

@@ -1,9 +1,25 @@
 package codex
 
-var ModelList = []string{
+import (
+	"github.com/QuantumNous/new-api/setting/ratio_setting"
+	"github.com/samber/lo"
+)
+
+var baseModelList = []string{
 	"gpt-5", "gpt-5-codex", "gpt-5-codex-mini",
 	"gpt-5.1", "gpt-5.1-codex", "gpt-5.1-codex-max", "gpt-5.1-codex-mini",
 	"gpt-5.2", "gpt-5.2-codex",
 }
 
+var ModelList = withCompactModelSuffix(baseModelList)
+
 const ChannelName = "codex"
+
+func withCompactModelSuffix(models []string) []string {
+	out := make([]string, 0, len(models)*2)
+	out = append(out, models...)
+	out = append(out, lo.Map(models, func(model string, _ int) string {
+		return ratio_setting.WithCompactModelSuffix(model)
+	})...)
+	return lo.Uniq(out)
+}

relay/channel/openai/adaptor.go

@@ -620,6 +620,8 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, info *relaycom
 		} else {
 			usage, err = OaiResponsesHandler(c, info, resp)
 		}
+	case relayconstant.RelayModeResponsesCompact:
+		usage, err = OaiResponsesCompactionHandler(c, resp)
 	default:
 		if info.IsStream {
 			usage, err = OaiStreamHandler(c, info, resp)

relay/channel/openai/relay_responses_compact.go

@@ -0,0 +1,44 @@
+package openai
+
+import (
+	"io"
+	"net/http"
+
+	"github.com/QuantumNous/new-api/common"
+	"github.com/QuantumNous/new-api/dto"
+	"github.com/QuantumNous/new-api/service"
+	"github.com/QuantumNous/new-api/types"
+
+	"github.com/gin-gonic/gin"
+)
+
+func OaiResponsesCompactionHandler(c *gin.Context, resp *http.Response) (*dto.Usage, *types.NewAPIError) {
+	defer service.CloseResponseBodyGracefully(resp)
+
+	responseBody, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return nil, types.NewOpenAIError(err, types.ErrorCodeReadResponseBodyFailed, http.StatusInternalServerError)
+	}
+
+	var compactResp dto.OpenAIResponsesCompactionResponse
+	if err := common.Unmarshal(responseBody, &compactResp); err != nil {
+		return nil, types.NewOpenAIError(err, types.ErrorCodeBadResponseBody, http.StatusInternalServerError)
+	}
+	if oaiError := compactResp.GetOpenAIError(); oaiError != nil && oaiError.Type != "" {
+		return nil, types.WithOpenAIError(*oaiError, resp.StatusCode)
+	}
+
+	service.IOCopyBytesGracefully(c, resp, responseBody)
+
+	usage := dto.Usage{}
+	if compactResp.Usage != nil {
+		usage.PromptTokens = compactResp.Usage.InputTokens
+		usage.CompletionTokens = compactResp.Usage.OutputTokens
+		usage.TotalTokens = compactResp.Usage.TotalTokens
+		if compactResp.Usage.InputTokensDetails != nil {
+			usage.PromptTokensDetails.CachedTokens = compactResp.Usage.InputTokensDetails.CachedTokens
+		}
+	}
+
+	return &usage, nil
+}

relay/common/relay_info.go

@@ -481,6 +481,11 @@ func GenRelayInfo(c *gin.Context, relayFormat types.RelayFormat, request dto.Req
 			break
 		}
 		err = errors.New("request is not a OpenAIResponsesRequest")
+	case types.RelayFormatOpenAIResponsesCompaction:
+		if request, ok := request.(*dto.OpenAIResponsesCompactionRequest); ok {
+			return GenRelayInfoResponsesCompaction(c, request), nil
+		}
+		return nil, errors.New("request is not a OpenAIResponsesCompactionRequest")
 	case types.RelayFormatTask:
 		info = genBaseRelayInfo(c, nil)
 	case types.RelayFormatMjProxy:
@@ -531,6 +536,15 @@ func (info *RelayInfo) AppendRequestConversion(format types.RelayFormat) {
 	info.RequestConversionChain = append(info.RequestConversionChain, format)
 }
 
+func GenRelayInfoResponsesCompaction(c *gin.Context, request *dto.OpenAIResponsesCompactionRequest) *RelayInfo {
+	info := genBaseRelayInfo(c, request)
+	if info.RelayMode == relayconstant.RelayModeUnknown {
+		info.RelayMode = relayconstant.RelayModeResponsesCompact
+	}
+	info.RelayFormat = types.RelayFormatOpenAIResponsesCompaction
+	return info
+}
+
 //func (info *RelayInfo) SetPromptTokens(promptTokens int) {
 //	info.promptTokens = promptTokens
 //}

relay/constant/relay_mode.go

@@ -50,6 +50,8 @@ const (
 	RelayModeRealtime
 
 	RelayModeGemini
+
+	RelayModeResponsesCompact
 )
 
 func Path2RelayMode(path string) int {
@@ -70,6 +72,8 @@ func Path2RelayMode(path string) int {
 		relayMode = RelayModeImagesEdits
 	} else if strings.HasPrefix(path, "/v1/edits") {
 		relayMode = RelayModeEdits
+	} else if strings.HasPrefix(path, "/v1/responses/compact") {
+		relayMode = RelayModeResponsesCompact
 	} else if strings.HasPrefix(path, "/v1/responses") {
 		relayMode = RelayModeResponses
 	} else if strings.HasPrefix(path, "/v1/audio/speech") {

relay/helper/model_mapped.go

@@ -4,13 +4,27 @@ import (
 	"encoding/json"
 	"errors"
 	"fmt"
+	"strings"
 
 	"github.com/QuantumNous/new-api/dto"
 	"github.com/QuantumNous/new-api/relay/common"
+	relayconstant "github.com/QuantumNous/new-api/relay/constant"
+	"github.com/QuantumNous/new-api/setting/ratio_setting"
 	"github.com/gin-gonic/gin"
 )
 
 func ModelMappedHelper(c *gin.Context, info *common.RelayInfo, request dto.Request) error {
+	if info.ChannelMeta == nil {
+		info.ChannelMeta = &common.ChannelMeta{}
+	}
+
+	isResponsesCompact := info.RelayMode == relayconstant.RelayModeResponsesCompact
+	originModelName := info.OriginModelName
+	mappingModelName := originModelName
+	if isResponsesCompact && strings.HasSuffix(originModelName, ratio_setting.CompactModelSuffix) {
+		mappingModelName = strings.TrimSuffix(originModelName, ratio_setting.CompactModelSuffix)
+	}
+
 	// map model name
 	modelMapping := c.GetString("model_mapping")
 	if modelMapping != "" && modelMapping != "{}" {
@@ -21,7 +35,7 @@ func ModelMappedHelper(c *gin.Context, info *common.RelayInfo, request dto.Reque
 		}
 
 		// 支持链式模型重定向，最终使用链尾的模型
-		currentModel := info.OriginModelName
+		currentModel := mappingModelName
 		visitedModels := map[string]bool{
 			currentModel: true,
 		}
@@ -51,6 +65,15 @@ func ModelMappedHelper(c *gin.Context, info *common.RelayInfo, request dto.Reque
 			info.UpstreamModelName = currentModel
 		}
 	}
+
+	if isResponsesCompact {
+		finalUpstreamModelName := mappingModelName
+		if info.IsModelMapped && info.UpstreamModelName != "" {
+			finalUpstreamModelName = info.UpstreamModelName
+		}
+		info.UpstreamModelName = finalUpstreamModelName
+		info.OriginModelName = ratio_setting.WithCompactModelSuffix(finalUpstreamModelName)
+	}
 	if request != nil {
 		request.SetModelName(info.UpstreamModelName)
 	}

relay/helper/valid_request.go

@@ -34,6 +34,8 @@ func GetAndValidateRequest(c *gin.Context, format types.RelayFormat) (request dt
 		request, err = GetAndValidateClaudeRequest(c)
 	case types.RelayFormatOpenAIResponses:
 		request, err = GetAndValidateResponsesRequest(c)
+	case types.RelayFormatOpenAIResponsesCompaction:
+		request, err = GetAndValidateResponsesCompactionRequest(c)
 
 	case types.RelayFormatOpenAIImage:
 		request, err = GetAndValidOpenAIImageRequest(c, relayMode)
@@ -125,6 +127,17 @@ func GetAndValidateResponsesRequest(c *gin.Context) (*dto.OpenAIResponsesRequest
 	return request, nil
 }
 
+func GetAndValidateResponsesCompactionRequest(c *gin.Context) (*dto.OpenAIResponsesCompactionRequest, error) {
+	request := &dto.OpenAIResponsesCompactionRequest{}
+	if err := common.UnmarshalBodyReusable(c, request); err != nil {
+		return nil, err
+	}
+	if request.Model == "" {
+		return nil, errors.New("model is required")
+	}
+	return request, nil
+}
+
 func GetAndValidOpenAIImageRequest(c *gin.Context, relayMode int) (*dto.ImageRequest, error) {
 	imageRequest := &dto.ImageRequest{}
 

relay/responses_handler.go

@@ -8,8 +8,10 @@ import (
 	"strings"
 
 	"github.com/QuantumNous/new-api/common"
+	appconstant "github.com/QuantumNous/new-api/constant"
 	"github.com/QuantumNous/new-api/dto"
 	relaycommon "github.com/QuantumNous/new-api/relay/common"
+	relayconstant "github.com/QuantumNous/new-api/relay/constant"
 	"github.com/QuantumNous/new-api/relay/helper"
 	"github.com/QuantumNous/new-api/service"
 	"github.com/QuantumNous/new-api/setting/model_setting"
@@ -20,10 +22,37 @@ import (
 
 func ResponsesHelper(c *gin.Context, info *relaycommon.RelayInfo) (newAPIError *types.NewAPIError) {
 	info.InitChannelMeta(c)
+	if info.RelayMode == relayconstant.RelayModeResponsesCompact {
+		switch info.ApiType {
+		case appconstant.APITypeOpenAI, appconstant.APITypeCodex:
+		default:
+			return types.NewErrorWithStatusCode(
+				fmt.Errorf("unsupported endpoint %q for api type %d", "/v1/responses/compact", info.ApiType),
+				types.ErrorCodeInvalidRequest,
+				http.StatusBadRequest,
+				types.ErrOptionWithSkipRetry(),
+			)
+		}
+	}
 
-	responsesReq, ok := info.Request.(*dto.OpenAIResponsesRequest)
-	if !ok {
-		return types.NewErrorWithStatusCode(fmt.Errorf("invalid request type, expected dto.OpenAIResponsesRequest, got %T", info.Request), types.ErrorCodeInvalidRequest, http.StatusBadRequest, types.ErrOptionWithSkipRetry())
+	var responsesReq *dto.OpenAIResponsesRequest
+	switch req := info.Request.(type) {
+	case *dto.OpenAIResponsesRequest:
+		responsesReq = req
+	case *dto.OpenAIResponsesCompactionRequest:
+		responsesReq = &dto.OpenAIResponsesRequest{
+			Model:              req.Model,
+			Input:              req.Input,
+			Instructions:       req.Instructions,
+			PreviousResponseID: req.PreviousResponseID,
+		}
+	default:
+		return types.NewErrorWithStatusCode(
+			fmt.Errorf("invalid request type, expected dto.OpenAIResponsesRequest or dto.OpenAIResponsesCompactionRequest, got %T", info.Request),
+			types.ErrorCodeInvalidRequest,
+			http.StatusBadRequest,
+			types.ErrOptionWithSkipRetry(),
+		)
 	}
 
 	request, err := common.DeepCopy(responsesReq)
@@ -105,10 +134,28 @@ func ResponsesHelper(c *gin.Context, info *relaycommon.RelayInfo) (newAPIError *
 		return newAPIError
 	}
 
+	usageDto := usage.(*dto.Usage)
+	if info.RelayMode == relayconstant.RelayModeResponsesCompact {
+		originModelName := info.OriginModelName
+		originPriceData := info.PriceData
+
+		_, err := helper.ModelPriceHelper(c, info, info.GetEstimatePromptTokens(), &types.TokenCountMeta{})
+		if err != nil {
+			info.OriginModelName = originModelName
+			info.PriceData = originPriceData
+			return types.NewError(err, types.ErrorCodeModelPriceError, types.ErrOptionWithSkipRetry())
+		}
+		postConsumeQuota(c, info, usageDto)
+
+		info.OriginModelName = originModelName
+		info.PriceData = originPriceData
+		return nil
+	}
+
 	if strings.HasPrefix(info.OriginModelName, "gpt-4o-audio") {
-		service.PostAudioConsumeQuota(c, info, usage.(*dto.Usage), "")
+		service.PostAudioConsumeQuota(c, info, usageDto, "")
 	} else {
-		postConsumeQuota(c, info, usage.(*dto.Usage))
+		postConsumeQuota(c, info, usageDto)
 	}
 	return nil
 }