test(responses): add comprehensive tests for SSE event ordering and response transformations

2026-02-03 04:50:52 +08:00 · 2026-01-21 07:08:59 +08:00
parent c6fa1d0e67
commit 523b41ccd2
2 changed files with 526 additions and 69 deletions
--- a/internal/translator/gemini/openai/responses/gemini_openai-responses_response.go
+++ b/internal/translator/gemini/openai/responses/gemini_openai-responses_response.go
@@ -20,6 +20,7 @@ type geminiToResponsesState struct {

 	// message aggregation
 	MsgOpened    bool
+	MsgClosed    bool
 	MsgIndex     int
 	CurrentMsgID string
 	TextBuf      strings.Builder
@@ -29,6 +30,7 @@ type geminiToResponsesState struct {
 	ReasoningOpened bool
 	ReasoningIndex  int
 	ReasoningItemID string
+	ReasoningEnc    string
 	ReasoningBuf    strings.Builder
 	ReasoningClosed bool

@@ -37,6 +39,7 @@ type geminiToResponsesState struct {
 	FuncArgsBuf map[int]*strings.Builder
 	FuncNames   map[int]string
 	FuncCallIDs map[int]string
+	FuncDone    map[int]bool
 }

 // responseIDCounter provides a process-wide unique counter for synthesized response identifiers.
@@ -45,6 +48,39 @@ var responseIDCounter uint64
 // funcCallIDCounter provides a process-wide unique counter for function call identifiers.
 var funcCallIDCounter uint64

+func pickRequestJSON(originalRequestRawJSON, requestRawJSON []byte) []byte {
+	if len(originalRequestRawJSON) > 0 && gjson.ValidBytes(originalRequestRawJSON) {
+		return originalRequestRawJSON
+	}
+	if len(requestRawJSON) > 0 && gjson.ValidBytes(requestRawJSON) {
+		return requestRawJSON
+	}
+	return nil
+}
+
+func unwrapRequestRoot(root gjson.Result) gjson.Result {
+	req := root.Get("request")
+	if !req.Exists() {
+		return root
+	}
+	if req.Get("model").Exists() || req.Get("input").Exists() || req.Get("instructions").Exists() {
+		return req
+	}
+	return root
+}
+
+func unwrapGeminiResponseRoot(root gjson.Result) gjson.Result {
+	resp := root.Get("response")
+	if !resp.Exists() {
+		return root
+	}
+	// Vertex-style Gemini responses wrap the actual payload in a "response" object.
+	if resp.Get("candidates").Exists() || resp.Get("responseId").Exists() || resp.Get("usageMetadata").Exists() {
+		return resp
+	}
+	return root
+}
+
 func emitEvent(event string, payload string) string {
 	return fmt.Sprintf("event: %s\ndata: %s", event, payload)
 }
@@ -56,18 +92,37 @@ func ConvertGeminiResponseToOpenAIResponses(_ context.Context, modelName string,
 			FuncArgsBuf: make(map[int]*strings.Builder),
 			FuncNames:   make(map[int]string),
 			FuncCallIDs: make(map[int]string),
+			FuncDone:    make(map[int]bool),
 		}
 	}
 	st := (*param).(*geminiToResponsesState)
+	if st.FuncArgsBuf == nil {
+		st.FuncArgsBuf = make(map[int]*strings.Builder)
+	}
+	if st.FuncNames == nil {
+		st.FuncNames = make(map[int]string)
+	}
+	if st.FuncCallIDs == nil {
+		st.FuncCallIDs = make(map[int]string)
+	}
+	if st.FuncDone == nil {
+		st.FuncDone = make(map[int]bool)
+	}

 	if bytes.HasPrefix(rawJSON, []byte("data:")) {
 		rawJSON = bytes.TrimSpace(rawJSON[5:])
 	}

+	rawJSON = bytes.TrimSpace(rawJSON)
+	if len(rawJSON) == 0 || bytes.Equal(rawJSON, []byte("[DONE]")) {
+		return []string{}
+	}
+
 	root := gjson.ParseBytes(rawJSON)
 	if !root.Exists() {
 		return []string{}
 	}
+	root = unwrapGeminiResponseRoot(root)

 	var out []string
 	nextSeq := func() int { st.Seq++; return st.Seq }
@@ -98,19 +153,54 @@ func ConvertGeminiResponseToOpenAIResponses(_ context.Context, modelName string,
 		itemDone, _ = sjson.Set(itemDone, "sequence_number", nextSeq())
 		itemDone, _ = sjson.Set(itemDone, "item.id", st.ReasoningItemID)
 		itemDone, _ = sjson.Set(itemDone, "output_index", st.ReasoningIndex)
+		itemDone, _ = sjson.Set(itemDone, "item.encrypted_content", st.ReasoningEnc)
 		itemDone, _ = sjson.Set(itemDone, "item.summary.0.text", full)
 		out = append(out, emitEvent("response.output_item.done", itemDone))

 		st.ReasoningClosed = true
 	}

+	// Helper to finalize the assistant message in correct order.
+	// It emits response.output_text.done, response.content_part.done,
+	// and response.output_item.done exactly once.
+	finalizeMessage := func() {
+		if !st.MsgOpened || st.MsgClosed {
+			return
+		}
+		fullText := st.ItemTextBuf.String()
+		done := `{"type":"response.output_text.done","sequence_number":0,"item_id":"","output_index":0,"content_index":0,"text":"","logprobs":[]}`
+		done, _ = sjson.Set(done, "sequence_number", nextSeq())
+		done, _ = sjson.Set(done, "item_id", st.CurrentMsgID)
+		done, _ = sjson.Set(done, "output_index", st.MsgIndex)
+		done, _ = sjson.Set(done, "text", fullText)
+		out = append(out, emitEvent("response.output_text.done", done))
+		partDone := `{"type":"response.content_part.done","sequence_number":0,"item_id":"","output_index":0,"content_index":0,"part":{"type":"output_text","annotations":[],"logprobs":[],"text":""}}`
+		partDone, _ = sjson.Set(partDone, "sequence_number", nextSeq())
+		partDone, _ = sjson.Set(partDone, "item_id", st.CurrentMsgID)
+		partDone, _ = sjson.Set(partDone, "output_index", st.MsgIndex)
+		partDone, _ = sjson.Set(partDone, "part.text", fullText)
+		out = append(out, emitEvent("response.content_part.done", partDone))
+		final := `{"type":"response.output_item.done","sequence_number":0,"output_index":0,"item":{"id":"","type":"message","status":"completed","content":[{"type":"output_text","text":""}],"role":"assistant"}}`
+		final, _ = sjson.Set(final, "sequence_number", nextSeq())
+		final, _ = sjson.Set(final, "output_index", st.MsgIndex)
+		final, _ = sjson.Set(final, "item.id", st.CurrentMsgID)
+		final, _ = sjson.Set(final, "item.content.0.text", fullText)
+		out = append(out, emitEvent("response.output_item.done", final))
+
+		st.MsgClosed = true
+	}
+
 	// Initialize per-response fields and emit created/in_progress once
 	if !st.Started {
-		if v := root.Get("responseId"); v.Exists() {
-			st.ResponseID = v.String()
+		st.ResponseID = root.Get("responseId").String()
+		if st.ResponseID == "" {
+			st.ResponseID = fmt.Sprintf("resp_%x_%d", time.Now().UnixNano(), atomic.AddUint64(&responseIDCounter, 1))
+		}
+		if !strings.HasPrefix(st.ResponseID, "resp_") {
+			st.ResponseID = fmt.Sprintf("resp_%s", st.ResponseID)
 		}
 		if v := root.Get("createTime"); v.Exists() {
-			if t, err := time.Parse(time.RFC3339Nano, v.String()); err == nil {
+			if t, errParseCreateTime := time.Parse(time.RFC3339Nano, v.String()); errParseCreateTime == nil {
 				st.CreatedAt = t.Unix()
 			}
 		}
@@ -143,15 +233,21 @@ func ConvertGeminiResponseToOpenAIResponses(_ context.Context, modelName string,
 					// Ignore any late thought chunks after reasoning is finalized.
 					return true
 				}
+				if sig := part.Get("thoughtSignature"); sig.Exists() && sig.String() != "" && sig.String() != geminiResponsesThoughtSignature {
+					st.ReasoningEnc = sig.String()
+				} else if sig = part.Get("thought_signature"); sig.Exists() && sig.String() != "" && sig.String() != geminiResponsesThoughtSignature {
+					st.ReasoningEnc = sig.String()
+				}
 				if !st.ReasoningOpened {
 					st.ReasoningOpened = true
 					st.ReasoningIndex = st.NextIndex
 					st.NextIndex++
 					st.ReasoningItemID = fmt.Sprintf("rs_%s_%d", st.ResponseID, st.ReasoningIndex)
-					item := `{"type":"response.output_item.added","sequence_number":0,"output_index":0,"item":{"id":"","type":"reasoning","status":"in_progress","summary":[]}}`
+					item := `{"type":"response.output_item.added","sequence_number":0,"output_index":0,"item":{"id":"","type":"reasoning","status":"in_progress","encrypted_content":"","summary":[]}}`
 					item, _ = sjson.Set(item, "sequence_number", nextSeq())
 					item, _ = sjson.Set(item, "output_index", st.ReasoningIndex)
 					item, _ = sjson.Set(item, "item.id", st.ReasoningItemID)
+					item, _ = sjson.Set(item, "item.encrypted_content", st.ReasoningEnc)
 					out = append(out, emitEvent("response.output_item.added", item))
 					partAdded := `{"type":"response.reasoning_summary_part.added","sequence_number":0,"item_id":"","output_index":0,"summary_index":0,"part":{"type":"summary_text","text":""}}`
 					partAdded, _ = sjson.Set(partAdded, "sequence_number", nextSeq())
@@ -191,9 +287,9 @@ func ConvertGeminiResponseToOpenAIResponses(_ context.Context, modelName string,
 					partAdded, _ = sjson.Set(partAdded, "output_index", st.MsgIndex)
 					out = append(out, emitEvent("response.content_part.added", partAdded))
 					st.ItemTextBuf.Reset()
-					st.ItemTextBuf.WriteString(t.String())
 				}
 				st.TextBuf.WriteString(t.String())
+				st.ItemTextBuf.WriteString(t.String())
 				msg := `{"type":"response.output_text.delta","sequence_number":0,"item_id":"","output_index":0,"content_index":0,"delta":"","logprobs":[]}`
 				msg, _ = sjson.Set(msg, "sequence_number", nextSeq())
 				msg, _ = sjson.Set(msg, "item_id", st.CurrentMsgID)
@@ -205,8 +301,10 @@ func ConvertGeminiResponseToOpenAIResponses(_ context.Context, modelName string,

 			// Function call
 			if fc := part.Get("functionCall"); fc.Exists() {
-				// Before emitting function-call outputs, finalize reasoning if open.
+				// Before emitting function-call outputs, finalize reasoning and the message (if open).
+				// Responses streaming requires message done events before the next output_item.added.
 				finalizeReasoning()
+				finalizeMessage()
 				name := fc.Get("name").String()
 				idx := st.NextIndex
 				st.NextIndex++
@@ -219,6 +317,14 @@ func ConvertGeminiResponseToOpenAIResponses(_ context.Context, modelName string,
 				}
 				st.FuncNames[idx] = name

+				argsJSON := "{}"
+				if args := fc.Get("args"); args.Exists() {
+					argsJSON = args.Raw
+				}
+				if st.FuncArgsBuf[idx].Len() == 0 && argsJSON != "" {
+					st.FuncArgsBuf[idx].WriteString(argsJSON)
+				}
+
 				// Emit item.added for function call
 				item := `{"type":"response.output_item.added","sequence_number":0,"output_index":0,"item":{"id":"","type":"function_call","status":"in_progress","arguments":"","call_id":"","name":""}}`
 				item, _ = sjson.Set(item, "sequence_number", nextSeq())
@@ -228,10 +334,9 @@ func ConvertGeminiResponseToOpenAIResponses(_ context.Context, modelName string,
 				item, _ = sjson.Set(item, "item.name", name)
 				out = append(out, emitEvent("response.output_item.added", item))

-				// Emit arguments delta (full args in one chunk)
-				if args := fc.Get("args"); args.Exists() {
-					argsJSON := args.Raw
-					st.FuncArgsBuf[idx].WriteString(argsJSON)
+				// Emit arguments delta (full args in one chunk).
+				// When Gemini omits args, emit "{}" to keep Responses streaming event order consistent.
+				if argsJSON != "" {
 					ad := `{"type":"response.function_call_arguments.delta","sequence_number":0,"item_id":"","output_index":0,"delta":""}`
 					ad, _ = sjson.Set(ad, "sequence_number", nextSeq())
 					ad, _ = sjson.Set(ad, "item_id", fmt.Sprintf("fc_%s", st.FuncCallIDs[idx]))
@@ -240,6 +345,27 @@ func ConvertGeminiResponseToOpenAIResponses(_ context.Context, modelName string,
 					out = append(out, emitEvent("response.function_call_arguments.delta", ad))
 				}

+				// Gemini emits the full function call payload at once, so we can finalize it immediately.
+				if !st.FuncDone[idx] {
+					fcDone := `{"type":"response.function_call_arguments.done","sequence_number":0,"item_id":"","output_index":0,"arguments":""}`
+					fcDone, _ = sjson.Set(fcDone, "sequence_number", nextSeq())
+					fcDone, _ = sjson.Set(fcDone, "item_id", fmt.Sprintf("fc_%s", st.FuncCallIDs[idx]))
+					fcDone, _ = sjson.Set(fcDone, "output_index", idx)
+					fcDone, _ = sjson.Set(fcDone, "arguments", argsJSON)
+					out = append(out, emitEvent("response.function_call_arguments.done", fcDone))
+
+					itemDone := `{"type":"response.output_item.done","sequence_number":0,"output_index":0,"item":{"id":"","type":"function_call","status":"completed","arguments":"","call_id":"","name":""}}`
+					itemDone, _ = sjson.Set(itemDone, "sequence_number", nextSeq())
+					itemDone, _ = sjson.Set(itemDone, "output_index", idx)
+					itemDone, _ = sjson.Set(itemDone, "item.id", fmt.Sprintf("fc_%s", st.FuncCallIDs[idx]))
+					itemDone, _ = sjson.Set(itemDone, "item.arguments", argsJSON)
+					itemDone, _ = sjson.Set(itemDone, "item.call_id", st.FuncCallIDs[idx])
+					itemDone, _ = sjson.Set(itemDone, "item.name", st.FuncNames[idx])
+					out = append(out, emitEvent("response.output_item.done", itemDone))
+
+					st.FuncDone[idx] = true
+				}
+
 				return true
 			}

@@ -251,28 +377,7 @@ func ConvertGeminiResponseToOpenAIResponses(_ context.Context, modelName string,
 	if fr := root.Get("candidates.0.finishReason"); fr.Exists() && fr.String() != "" {
 		// Finalize reasoning first to keep ordering tight with last delta
 		finalizeReasoning()
-		// Close message output if opened
-		if st.MsgOpened {
-			fullText := st.ItemTextBuf.String()
-			done := `{"type":"response.output_text.done","sequence_number":0,"item_id":"","output_index":0,"content_index":0,"text":"","logprobs":[]}`
-			done, _ = sjson.Set(done, "sequence_number", nextSeq())
-			done, _ = sjson.Set(done, "item_id", st.CurrentMsgID)
-			done, _ = sjson.Set(done, "output_index", st.MsgIndex)
-			done, _ = sjson.Set(done, "text", fullText)
-			out = append(out, emitEvent("response.output_text.done", done))
-			partDone := `{"type":"response.content_part.done","sequence_number":0,"item_id":"","output_index":0,"content_index":0,"part":{"type":"output_text","annotations":[],"logprobs":[],"text":""}}`
-			partDone, _ = sjson.Set(partDone, "sequence_number", nextSeq())
-			partDone, _ = sjson.Set(partDone, "item_id", st.CurrentMsgID)
-			partDone, _ = sjson.Set(partDone, "output_index", st.MsgIndex)
-			partDone, _ = sjson.Set(partDone, "part.text", fullText)
-			out = append(out, emitEvent("response.content_part.done", partDone))
-			final := `{"type":"response.output_item.done","sequence_number":0,"output_index":0,"item":{"id":"","type":"message","status":"completed","content":[{"type":"output_text","text":""}],"role":"assistant"}}`
-			final, _ = sjson.Set(final, "sequence_number", nextSeq())
-			final, _ = sjson.Set(final, "output_index", st.MsgIndex)
-			final, _ = sjson.Set(final, "item.id", st.CurrentMsgID)
-			final, _ = sjson.Set(final, "item.content.0.text", fullText)
-			out = append(out, emitEvent("response.output_item.done", final))
-		}
+		finalizeMessage()

 		// Close function calls
 		if len(st.FuncArgsBuf) > 0 {
@@ -289,6 +394,9 @@ func ConvertGeminiResponseToOpenAIResponses(_ context.Context, modelName string,
 				}
 			}
 			for _, idx := range idxs {
+				if st.FuncDone[idx] {
+					continue
+				}
 				args := "{}"
 				if b := st.FuncArgsBuf[idx]; b != nil && b.Len() > 0 {
 					args = b.String()
@@ -308,6 +416,8 @@ func ConvertGeminiResponseToOpenAIResponses(_ context.Context, modelName string,
 				itemDone, _ = sjson.Set(itemDone, "item.call_id", st.FuncCallIDs[idx])
 				itemDone, _ = sjson.Set(itemDone, "item.name", st.FuncNames[idx])
 				out = append(out, emitEvent("response.output_item.done", itemDone))
+
+				st.FuncDone[idx] = true
 			}
 		}

@@ -319,8 +429,8 @@ func ConvertGeminiResponseToOpenAIResponses(_ context.Context, modelName string,
 		completed, _ = sjson.Set(completed, "response.id", st.ResponseID)
 		completed, _ = sjson.Set(completed, "response.created_at", st.CreatedAt)

-		if requestRawJSON != nil {
-			req := gjson.ParseBytes(requestRawJSON)
+		if reqJSON := pickRequestJSON(originalRequestRawJSON, requestRawJSON); len(reqJSON) > 0 {
+			req := unwrapRequestRoot(gjson.ParseBytes(reqJSON))
 			if v := req.Get("instructions"); v.Exists() {
 				completed, _ = sjson.Set(completed, "response.instructions", v.String())
 			}
@@ -383,41 +493,34 @@ func ConvertGeminiResponseToOpenAIResponses(_ context.Context, modelName string,
 			}
 		}

-		// Compose outputs in encountered order: reasoning, message, function_calls
+		// Compose outputs in output_index order.
 		outputsWrapper := `{"arr":[]}`
-		if st.ReasoningOpened {
-			item := `{"id":"","type":"reasoning","summary":[{"type":"summary_text","text":""}]}`
-			item, _ = sjson.Set(item, "id", st.ReasoningItemID)
-			item, _ = sjson.Set(item, "summary.0.text", st.ReasoningBuf.String())
-			outputsWrapper, _ = sjson.SetRaw(outputsWrapper, "arr.-1", item)
-		}
-		if st.MsgOpened {
-			item := `{"id":"","type":"message","status":"completed","content":[{"type":"output_text","annotations":[],"logprobs":[],"text":""}],"role":"assistant"}`
-			item, _ = sjson.Set(item, "id", st.CurrentMsgID)
-			item, _ = sjson.Set(item, "content.0.text", st.TextBuf.String())
-			outputsWrapper, _ = sjson.SetRaw(outputsWrapper, "arr.-1", item)
-		}
-		if len(st.FuncArgsBuf) > 0 {
-			idxs := make([]int, 0, len(st.FuncArgsBuf))
-			for idx := range st.FuncArgsBuf {
-				idxs = append(idxs, idx)
+		for idx := 0; idx < st.NextIndex; idx++ {
+			if st.ReasoningOpened && idx == st.ReasoningIndex {
+				item := `{"id":"","type":"reasoning","encrypted_content":"","summary":[{"type":"summary_text","text":""}]}`
+				item, _ = sjson.Set(item, "id", st.ReasoningItemID)
+				item, _ = sjson.Set(item, "encrypted_content", st.ReasoningEnc)
+				item, _ = sjson.Set(item, "summary.0.text", st.ReasoningBuf.String())
+				outputsWrapper, _ = sjson.SetRaw(outputsWrapper, "arr.-1", item)
+				continue
 			}
-			for i := 0; i < len(idxs); i++ {
-				for j := i + 1; j < len(idxs); j++ {
-					if idxs[j] < idxs[i] {
-						idxs[i], idxs[j] = idxs[j], idxs[i]
-					}
-				}
+			if st.MsgOpened && idx == st.MsgIndex {
+				item := `{"id":"","type":"message","status":"completed","content":[{"type":"output_text","annotations":[],"logprobs":[],"text":""}],"role":"assistant"}`
+				item, _ = sjson.Set(item, "id", st.CurrentMsgID)
+				item, _ = sjson.Set(item, "content.0.text", st.TextBuf.String())
+				outputsWrapper, _ = sjson.SetRaw(outputsWrapper, "arr.-1", item)
+				continue
 			}
-			for _, idx := range idxs {
-				args := ""
-				if b := st.FuncArgsBuf[idx]; b != nil {
+
+			if callID, ok := st.FuncCallIDs[idx]; ok && callID != "" {
+				args := "{}"
+				if b := st.FuncArgsBuf[idx]; b != nil && b.Len() > 0 {
 					args = b.String()
 				}
 				item := `{"id":"","type":"function_call","status":"completed","arguments":"","call_id":"","name":""}`
-				item, _ = sjson.Set(item, "id", fmt.Sprintf("fc_%s", st.FuncCallIDs[idx]))
+				item, _ = sjson.Set(item, "id", fmt.Sprintf("fc_%s", callID))
 				item, _ = sjson.Set(item, "arguments", args)
-				item, _ = sjson.Set(item, "call_id", st.FuncCallIDs[idx])
+				item, _ = sjson.Set(item, "call_id", callID)
 				item, _ = sjson.Set(item, "name", st.FuncNames[idx])
 				outputsWrapper, _ = sjson.SetRaw(outputsWrapper, "arr.-1", item)
 			}
@@ -431,8 +534,8 @@ func ConvertGeminiResponseToOpenAIResponses(_ context.Context, modelName string,
 			// input tokens = prompt + thoughts
 			input := um.Get("promptTokenCount").Int() + um.Get("thoughtsTokenCount").Int()
 			completed, _ = sjson.Set(completed, "response.usage.input_tokens", input)
-			// cached_tokens not provided by Gemini; default to 0 for structure compatibility
-			completed, _ = sjson.Set(completed, "response.usage.input_tokens_details.cached_tokens", 0)
+			// cached token details: align with OpenAI "cached_tokens" semantics.
+			completed, _ = sjson.Set(completed, "response.usage.input_tokens_details.cached_tokens", um.Get("cachedContentTokenCount").Int())
 			// output tokens
 			if v := um.Get("candidatesTokenCount"); v.Exists() {
 				completed, _ = sjson.Set(completed, "response.usage.output_tokens", v.Int())
@@ -460,6 +563,7 @@ func ConvertGeminiResponseToOpenAIResponses(_ context.Context, modelName string,
 // ConvertGeminiResponseToOpenAIResponsesNonStream aggregates Gemini response JSON into a single OpenAI Responses JSON object.
 func ConvertGeminiResponseToOpenAIResponsesNonStream(_ context.Context, _ string, originalRequestRawJSON, requestRawJSON, rawJSON []byte, _ *any) string {
 	root := gjson.ParseBytes(rawJSON)
+	root = unwrapGeminiResponseRoot(root)

 	// Base response scaffold
 	resp := `{"id":"","object":"response","created_at":0,"status":"completed","background":false,"error":null,"incomplete_details":null}`
@@ -478,15 +582,15 @@ func ConvertGeminiResponseToOpenAIResponsesNonStream(_ context.Context, _ string
 	// created_at: map from createTime if available
 	createdAt := time.Now().Unix()
 	if v := root.Get("createTime"); v.Exists() {
-		if t, err := time.Parse(time.RFC3339Nano, v.String()); err == nil {
+		if t, errParseCreateTime := time.Parse(time.RFC3339Nano, v.String()); errParseCreateTime == nil {
 			createdAt = t.Unix()
 		}
 	}
 	resp, _ = sjson.Set(resp, "created_at", createdAt)

 	// Echo request fields when present; fallback model from response modelVersion
-	if len(requestRawJSON) > 0 {
-		req := gjson.ParseBytes(requestRawJSON)
+	if reqJSON := pickRequestJSON(originalRequestRawJSON, requestRawJSON); len(reqJSON) > 0 {
+		req := unwrapRequestRoot(gjson.ParseBytes(reqJSON))
 		if v := req.Get("instructions"); v.Exists() {
 			resp, _ = sjson.Set(resp, "instructions", v.String())
 		}
@@ -636,8 +740,8 @@ func ConvertGeminiResponseToOpenAIResponsesNonStream(_ context.Context, _ string
 		// input tokens = prompt + thoughts
 		input := um.Get("promptTokenCount").Int() + um.Get("thoughtsTokenCount").Int()
 		resp, _ = sjson.Set(resp, "usage.input_tokens", input)
-		// cached_tokens not provided by Gemini; default to 0 for structure compatibility
-		resp, _ = sjson.Set(resp, "usage.input_tokens_details.cached_tokens", 0)
+		// cached token details: align with OpenAI "cached_tokens" semantics.
+		resp, _ = sjson.Set(resp, "usage.input_tokens_details.cached_tokens", um.Get("cachedContentTokenCount").Int())
 		// output tokens
 		if v := um.Get("candidatesTokenCount"); v.Exists() {
 			resp, _ = sjson.Set(resp, "usage.output_tokens", v.Int())
--- a/internal/translator/gemini/openai/responses/gemini_openai-responses_response_test.go
+++ b/internal/translator/gemini/openai/responses/gemini_openai-responses_response_test.go
@@ -0,0 +1,353 @@
+package responses
+
+import (
+	"context"
+	"strings"
+	"testing"
+
+	"github.com/tidwall/gjson"
+)
+
+func parseSSEEvent(t *testing.T, chunk string) (string, gjson.Result) {
+	t.Helper()
+
+	lines := strings.Split(chunk, "\n")
+	if len(lines) < 2 {
+		t.Fatalf("unexpected SSE chunk: %q", chunk)
+	}
+
+	event := strings.TrimSpace(strings.TrimPrefix(lines[0], "event:"))
+	dataLine := strings.TrimSpace(strings.TrimPrefix(lines[1], "data:"))
+	if !gjson.Valid(dataLine) {
+		t.Fatalf("invalid SSE data JSON: %q", dataLine)
+	}
+	return event, gjson.Parse(dataLine)
+}
+
+func TestConvertGeminiResponseToOpenAIResponses_UnwrapAndAggregateText(t *testing.T) {
+	// Vertex-style Gemini stream wraps the actual response payload under "response".
+	// This test ensures we unwrap and that output_text.done contains the full text.
+	in := []string{
+		`data: {"response":{"candidates":[{"content":{"role":"model","parts":[{"text":""}]}}],"usageMetadata":{"promptTokenCount":1,"candidatesTokenCount":1,"totalTokenCount":2,"cachedContentTokenCount":0},"modelVersion":"test-model","responseId":"req_vrtx_1"},"traceId":"t1"}`,
+		`data: {"response":{"candidates":[{"content":{"role":"model","parts":[{"text":"让"}]}}],"usageMetadata":{"promptTokenCount":1,"candidatesTokenCount":1,"totalTokenCount":2,"cachedContentTokenCount":0},"modelVersion":"test-model","responseId":"req_vrtx_1"},"traceId":"t1"}`,
+		`data: {"response":{"candidates":[{"content":{"role":"model","parts":[{"text":"我先"}]}}],"usageMetadata":{"promptTokenCount":1,"candidatesTokenCount":1,"totalTokenCount":2,"cachedContentTokenCount":0},"modelVersion":"test-model","responseId":"req_vrtx_1"},"traceId":"t1"}`,
+		`data: {"response":{"candidates":[{"content":{"role":"model","parts":[{"text":"了解"}]}}],"usageMetadata":{"promptTokenCount":1,"candidatesTokenCount":1,"totalTokenCount":2,"cachedContentTokenCount":0},"modelVersion":"test-model","responseId":"req_vrtx_1"},"traceId":"t1"}`,
+		`data: {"response":{"candidates":[{"content":{"role":"model","parts":[{"functionCall":{"name":"mcp__serena__list_dir","args":{"recursive":false,"relative_path":"internal"},"id":"toolu_1"}}]}}],"usageMetadata":{"promptTokenCount":1,"candidatesTokenCount":1,"totalTokenCount":2,"cachedContentTokenCount":0},"modelVersion":"test-model","responseId":"req_vrtx_1"},"traceId":"t1"}`,
+		`data: {"response":{"candidates":[{"content":{"role":"model","parts":[{"text":""}]},"finishReason":"STOP"}],"usageMetadata":{"promptTokenCount":10,"candidatesTokenCount":5,"totalTokenCount":15,"cachedContentTokenCount":2},"modelVersion":"test-model","responseId":"req_vrtx_1"},"traceId":"t1"}`,
+	}
+
+	originalReq := []byte(`{"instructions":"test instructions","model":"gpt-5","max_output_tokens":123}`)
+
+	var param any
+	var out []string
+	for _, line := range in {
+		out = append(out, ConvertGeminiResponseToOpenAIResponses(context.Background(), "test-model", originalReq, nil, []byte(line), &param)...)
+	}
+
+	var (
+		gotTextDone     bool
+		gotMessageDone  bool
+		gotResponseDone bool
+		gotFuncDone     bool
+
+		textDone     string
+		messageText  string
+		responseID   string
+		instructions string
+		cachedTokens int64
+
+		funcName string
+		funcArgs string
+
+		posTextDone    = -1
+		posPartDone    = -1
+		posMessageDone = -1
+		posFuncAdded   = -1
+	)
+
+	for i, chunk := range out {
+		ev, data := parseSSEEvent(t, chunk)
+		switch ev {
+		case "response.output_text.done":
+			gotTextDone = true
+			if posTextDone == -1 {
+				posTextDone = i
+			}
+			textDone = data.Get("text").String()
+		case "response.content_part.done":
+			if posPartDone == -1 {
+				posPartDone = i
+			}
+		case "response.output_item.done":
+			switch data.Get("item.type").String() {
+			case "message":
+				gotMessageDone = true
+				if posMessageDone == -1 {
+					posMessageDone = i
+				}
+				messageText = data.Get("item.content.0.text").String()
+			case "function_call":
+				gotFuncDone = true
+				funcName = data.Get("item.name").String()
+				funcArgs = data.Get("item.arguments").String()
+			}
+		case "response.output_item.added":
+			if data.Get("item.type").String() == "function_call" && posFuncAdded == -1 {
+				posFuncAdded = i
+			}
+		case "response.completed":
+			gotResponseDone = true
+			responseID = data.Get("response.id").String()
+			instructions = data.Get("response.instructions").String()
+			cachedTokens = data.Get("response.usage.input_tokens_details.cached_tokens").Int()
+		}
+	}
+
+	if !gotTextDone {
+		t.Fatalf("missing response.output_text.done event")
+	}
+	if posTextDone == -1 || posPartDone == -1 || posMessageDone == -1 || posFuncAdded == -1 {
+		t.Fatalf("missing ordering events: textDone=%d partDone=%d messageDone=%d funcAdded=%d", posTextDone, posPartDone, posMessageDone, posFuncAdded)
+	}
+	if !(posTextDone < posPartDone && posPartDone < posMessageDone && posMessageDone < posFuncAdded) {
+		t.Fatalf("unexpected message/function ordering: textDone=%d partDone=%d messageDone=%d funcAdded=%d", posTextDone, posPartDone, posMessageDone, posFuncAdded)
+	}
+	if !gotMessageDone {
+		t.Fatalf("missing message response.output_item.done event")
+	}
+	if !gotFuncDone {
+		t.Fatalf("missing function_call response.output_item.done event")
+	}
+	if !gotResponseDone {
+		t.Fatalf("missing response.completed event")
+	}
+
+	if textDone != "让我先了解" {
+		t.Fatalf("unexpected output_text.done text: got %q", textDone)
+	}
+	if messageText != "让我先了解" {
+		t.Fatalf("unexpected message done text: got %q", messageText)
+	}
+
+	if responseID != "resp_req_vrtx_1" {
+		t.Fatalf("unexpected response id: got %q", responseID)
+	}
+	if instructions != "test instructions" {
+		t.Fatalf("unexpected instructions echo: got %q", instructions)
+	}
+	if cachedTokens != 2 {
+		t.Fatalf("unexpected cached token count: got %d", cachedTokens)
+	}
+
+	if funcName != "mcp__serena__list_dir" {
+		t.Fatalf("unexpected function name: got %q", funcName)
+	}
+	if !gjson.Valid(funcArgs) {
+		t.Fatalf("invalid function arguments JSON: %q", funcArgs)
+	}
+	if gjson.Get(funcArgs, "recursive").Bool() != false {
+		t.Fatalf("unexpected recursive arg: %v", gjson.Get(funcArgs, "recursive").Value())
+	}
+	if gjson.Get(funcArgs, "relative_path").String() != "internal" {
+		t.Fatalf("unexpected relative_path arg: %q", gjson.Get(funcArgs, "relative_path").String())
+	}
+}
+
+func TestConvertGeminiResponseToOpenAIResponses_ReasoningEncryptedContent(t *testing.T) {
+	sig := "RXE0RENrZ0lDeEFDR0FJcVFOZDdjUzlleGFuRktRdFcvSzNyZ2MvWDNCcDQ4RmxSbGxOWUlOVU5kR1l1UHMrMGdkMVp0Vkg3ekdKU0g4YVljc2JjN3lNK0FrdGpTNUdqamI4T3Z0VVNETzdQd3pmcFhUOGl3U3hXUEJvTVFRQ09mWTFyMEtTWGZxUUlJakFqdmFGWk83RW1XRlBKckJVOVpkYzdDKw=="
+	in := []string{
+		`data: {"response":{"candidates":[{"content":{"role":"model","parts":[{"thought":true,"thoughtSignature":"` + sig + `","text":""}]}}],"modelVersion":"test-model","responseId":"req_vrtx_sig"},"traceId":"t1"}`,
+		`data: {"response":{"candidates":[{"content":{"role":"model","parts":[{"thought":true,"text":"a"}]}}],"modelVersion":"test-model","responseId":"req_vrtx_sig"},"traceId":"t1"}`,
+		`data: {"response":{"candidates":[{"content":{"role":"model","parts":[{"text":"hello"}]}}],"modelVersion":"test-model","responseId":"req_vrtx_sig"},"traceId":"t1"}`,
+		`data: {"response":{"candidates":[{"content":{"role":"model","parts":[{"text":""}]},"finishReason":"STOP"}],"modelVersion":"test-model","responseId":"req_vrtx_sig"},"traceId":"t1"}`,
+	}
+
+	var param any
+	var out []string
+	for _, line := range in {
+		out = append(out, ConvertGeminiResponseToOpenAIResponses(context.Background(), "test-model", nil, nil, []byte(line), &param)...)
+	}
+
+	var (
+		addedEnc string
+		doneEnc  string
+	)
+	for _, chunk := range out {
+		ev, data := parseSSEEvent(t, chunk)
+		switch ev {
+		case "response.output_item.added":
+			if data.Get("item.type").String() == "reasoning" {
+				addedEnc = data.Get("item.encrypted_content").String()
+			}
+		case "response.output_item.done":
+			if data.Get("item.type").String() == "reasoning" {
+				doneEnc = data.Get("item.encrypted_content").String()
+			}
+		}
+	}
+
+	if addedEnc != sig {
+		t.Fatalf("unexpected encrypted_content in response.output_item.added: got %q", addedEnc)
+	}
+	if doneEnc != sig {
+		t.Fatalf("unexpected encrypted_content in response.output_item.done: got %q", doneEnc)
+	}
+}
+
+func TestConvertGeminiResponseToOpenAIResponses_FunctionCallEventOrder(t *testing.T) {
+	in := []string{
+		`data: {"response":{"candidates":[{"content":{"role":"model","parts":[{"functionCall":{"name":"tool0"}}]}}],"modelVersion":"test-model","responseId":"req_vrtx_1"},"traceId":"t1"}`,
+		`data: {"response":{"candidates":[{"content":{"role":"model","parts":[{"functionCall":{"name":"tool1"}}]}}],"modelVersion":"test-model","responseId":"req_vrtx_1"},"traceId":"t1"}`,
+		`data: {"response":{"candidates":[{"content":{"role":"model","parts":[{"functionCall":{"name":"tool2","args":{"a":1}}}]}}],"modelVersion":"test-model","responseId":"req_vrtx_1"},"traceId":"t1"}`,
+		`data: {"response":{"candidates":[{"content":{"role":"model","parts":[{"text":""}]},"finishReason":"STOP"}],"usageMetadata":{"promptTokenCount":10,"candidatesTokenCount":5,"totalTokenCount":15,"cachedContentTokenCount":0},"modelVersion":"test-model","responseId":"req_vrtx_1"},"traceId":"t1"}`,
+	}
+
+	var param any
+	var out []string
+	for _, line := range in {
+		out = append(out, ConvertGeminiResponseToOpenAIResponses(context.Background(), "test-model", nil, nil, []byte(line), &param)...)
+	}
+
+	posAdded := []int{-1, -1, -1}
+	posArgsDelta := []int{-1, -1, -1}
+	posArgsDone := []int{-1, -1, -1}
+	posItemDone := []int{-1, -1, -1}
+	posCompleted := -1
+	deltaByIndex := map[int]string{}
+
+	for i, chunk := range out {
+		ev, data := parseSSEEvent(t, chunk)
+		switch ev {
+		case "response.output_item.added":
+			if data.Get("item.type").String() != "function_call" {
+				continue
+			}
+			idx := int(data.Get("output_index").Int())
+			if idx >= 0 && idx < len(posAdded) {
+				posAdded[idx] = i
+			}
+		case "response.function_call_arguments.delta":
+			idx := int(data.Get("output_index").Int())
+			if idx >= 0 && idx < len(posArgsDelta) {
+				posArgsDelta[idx] = i
+				deltaByIndex[idx] = data.Get("delta").String()
+			}
+		case "response.function_call_arguments.done":
+			idx := int(data.Get("output_index").Int())
+			if idx >= 0 && idx < len(posArgsDone) {
+				posArgsDone[idx] = i
+			}
+		case "response.output_item.done":
+			if data.Get("item.type").String() != "function_call" {
+				continue
+			}
+			idx := int(data.Get("output_index").Int())
+			if idx >= 0 && idx < len(posItemDone) {
+				posItemDone[idx] = i
+			}
+		case "response.completed":
+			posCompleted = i
+
+			output := data.Get("response.output")
+			if !output.Exists() || !output.IsArray() {
+				t.Fatalf("missing response.output in response.completed")
+			}
+			if len(output.Array()) != 3 {
+				t.Fatalf("unexpected response.output length: got %d", len(output.Array()))
+			}
+			if data.Get("response.output.0.name").String() != "tool0" || data.Get("response.output.0.arguments").String() != "{}" {
+				t.Fatalf("unexpected output[0]: %s", data.Get("response.output.0").Raw)
+			}
+			if data.Get("response.output.1.name").String() != "tool1" || data.Get("response.output.1.arguments").String() != "{}" {
+				t.Fatalf("unexpected output[1]: %s", data.Get("response.output.1").Raw)
+			}
+			if data.Get("response.output.2.name").String() != "tool2" {
+				t.Fatalf("unexpected output[2] name: %s", data.Get("response.output.2").Raw)
+			}
+			if !gjson.Valid(data.Get("response.output.2.arguments").String()) {
+				t.Fatalf("unexpected output[2] arguments: %q", data.Get("response.output.2.arguments").String())
+			}
+		}
+	}
+
+	if posCompleted == -1 {
+		t.Fatalf("missing response.completed event")
+	}
+	for idx := 0; idx < 3; idx++ {
+		if posAdded[idx] == -1 || posArgsDelta[idx] == -1 || posArgsDone[idx] == -1 || posItemDone[idx] == -1 {
+			t.Fatalf("missing function call events for output_index %d: added=%d argsDelta=%d argsDone=%d itemDone=%d", idx, posAdded[idx], posArgsDelta[idx], posArgsDone[idx], posItemDone[idx])
+		}
+		if !(posAdded[idx] < posArgsDelta[idx] && posArgsDelta[idx] < posArgsDone[idx] && posArgsDone[idx] < posItemDone[idx]) {
+			t.Fatalf("unexpected ordering for output_index %d: added=%d argsDelta=%d argsDone=%d itemDone=%d", idx, posAdded[idx], posArgsDelta[idx], posArgsDone[idx], posItemDone[idx])
+		}
+		if idx > 0 && !(posItemDone[idx-1] < posAdded[idx]) {
+			t.Fatalf("function call events overlap between %d and %d: prevDone=%d nextAdded=%d", idx-1, idx, posItemDone[idx-1], posAdded[idx])
+		}
+	}
+
+	if deltaByIndex[0] != "{}" {
+		t.Fatalf("unexpected delta for output_index 0: got %q", deltaByIndex[0])
+	}
+	if deltaByIndex[1] != "{}" {
+		t.Fatalf("unexpected delta for output_index 1: got %q", deltaByIndex[1])
+	}
+	if deltaByIndex[2] == "" || !gjson.Valid(deltaByIndex[2]) || gjson.Get(deltaByIndex[2], "a").Int() != 1 {
+		t.Fatalf("unexpected delta for output_index 2: got %q", deltaByIndex[2])
+	}
+	if !(posItemDone[2] < posCompleted) {
+		t.Fatalf("response.completed should be after last output_item.done: last=%d completed=%d", posItemDone[2], posCompleted)
+	}
+}
+
+func TestConvertGeminiResponseToOpenAIResponses_ResponseOutputOrdering(t *testing.T) {
+	in := []string{
+		`data: {"response":{"candidates":[{"content":{"role":"model","parts":[{"functionCall":{"name":"tool0","args":{"x":"y"}}}]}}],"modelVersion":"test-model","responseId":"req_vrtx_2"},"traceId":"t2"}`,
+		`data: {"response":{"candidates":[{"content":{"role":"model","parts":[{"text":"hi"}]}}],"modelVersion":"test-model","responseId":"req_vrtx_2"},"traceId":"t2"}`,
+		`data: {"response":{"candidates":[{"content":{"role":"model","parts":[{"text":""}]},"finishReason":"STOP"}],"usageMetadata":{"promptTokenCount":1,"candidatesTokenCount":1,"totalTokenCount":2,"cachedContentTokenCount":0},"modelVersion":"test-model","responseId":"req_vrtx_2"},"traceId":"t2"}`,
+	}
+
+	var param any
+	var out []string
+	for _, line := range in {
+		out = append(out, ConvertGeminiResponseToOpenAIResponses(context.Background(), "test-model", nil, nil, []byte(line), &param)...)
+	}
+
+	posFuncDone := -1
+	posMsgAdded := -1
+	posCompleted := -1
+
+	for i, chunk := range out {
+		ev, data := parseSSEEvent(t, chunk)
+		switch ev {
+		case "response.output_item.done":
+			if data.Get("item.type").String() == "function_call" && data.Get("output_index").Int() == 0 {
+				posFuncDone = i
+			}
+		case "response.output_item.added":
+			if data.Get("item.type").String() == "message" && data.Get("output_index").Int() == 1 {
+				posMsgAdded = i
+			}
+		case "response.completed":
+			posCompleted = i
+			if data.Get("response.output.0.type").String() != "function_call" {
+				t.Fatalf("expected response.output[0] to be function_call: %s", data.Get("response.output.0").Raw)
+			}
+			if data.Get("response.output.1.type").String() != "message" {
+				t.Fatalf("expected response.output[1] to be message: %s", data.Get("response.output.1").Raw)
+			}
+			if data.Get("response.output.1.content.0.text").String() != "hi" {
+				t.Fatalf("unexpected message text in response.output[1]: %s", data.Get("response.output.1").Raw)
+			}
+		}
+	}
+
+	if posFuncDone == -1 || posMsgAdded == -1 || posCompleted == -1 {
+		t.Fatalf("missing required events: funcDone=%d msgAdded=%d completed=%d", posFuncDone, posMsgAdded, posCompleted)
+	}
+	if !(posFuncDone < posMsgAdded) {
+		t.Fatalf("expected function_call to complete before message is added: funcDone=%d msgAdded=%d", posFuncDone, posMsgAdded)
+	}
+	if !(posMsgAdded < posCompleted) {
+		t.Fatalf("expected response.completed after message added: msgAdded=%d completed=%d", posMsgAdded, posCompleted)
+	}
+}