Merge branch 'main' into test

backend/internal/pkg/antigravity/gemini_types.go

@@ -155,6 +155,7 @@ type GeminiUsageMetadata struct {
 	CandidatesTokenCount    int `json:"candidatesTokenCount,omitempty"`
 	CachedContentTokenCount int `json:"cachedContentTokenCount,omitempty"`
 	TotalTokenCount         int `json:"totalTokenCount,omitempty"`
+	ThoughtsTokenCount      int `json:"thoughtsTokenCount,omitempty"` // thinking tokens（按输出价格计费）
 }
 
 // GeminiGroundingMetadata Gemini grounding 元数据（Web Search）

backend/internal/pkg/antigravity/request_transformer.go

@@ -64,6 +64,10 @@ const MaxTokensBudgetPadding = 1000
 // Gemini 2.5 Flash thinking budget 上限
 const Gemini25FlashThinkingBudgetLimit = 24576
 
+// 对于 Antigravity 的 Claude（budget-only）模型，该语义最终等价为 thinkingBudget=24576。
+// 这里复用相同数值以保持行为一致。
+const ClaudeAdaptiveHighThinkingBudgetTokens = Gemini25FlashThinkingBudgetLimit
+
 // ensureMaxTokensGreaterThanBudget 确保 max_tokens > budget_tokens
 // Claude API 要求启用 thinking 时，max_tokens 必须大于 thinking.budget_tokens
 // 返回调整后的 maxTokens 和是否进行了调整
@@ -96,7 +100,7 @@ func TransformClaudeToGeminiWithOptions(claudeReq *ClaudeRequest, projectID, map
 	}
 
 	// 检测是否启用 thinking
-	isThinkingEnabled := claudeReq.Thinking != nil && claudeReq.Thinking.Type == "enabled"
+	isThinkingEnabled := claudeReq.Thinking != nil && (claudeReq.Thinking.Type == "enabled" || claudeReq.Thinking.Type == "adaptive")
 
 	// 只有 Gemini 模型支持 dummy thought workaround
 	// Claude 模型通过 Vertex/Google API 需要有效的 thought signatures
@@ -198,8 +202,7 @@ type modelInfo struct {
 
 // modelInfoMap 模型前缀 → 模型信息映射
 // 只有在此映射表中的模型才会注入身份提示词
-// 注意：当前 claude-opus-4-6 会被映射到 claude-opus-4-5-thinking，
-// 但保留此条目以便后续 Antigravity 上游支持 4.6 时快速切换
+// 注意：模型映射逻辑在网关层完成；这里仅用于按模型前缀判断是否注入身份提示词。
 var modelInfoMap = map[string]modelInfo{
 	"claude-opus-4-5":   {DisplayName: "Claude Opus 4.5", CanonicalID: "claude-opus-4-5-20250929"},
 	"claude-opus-4-6":   {DisplayName: "Claude Opus 4.6", CanonicalID: "claude-opus-4-6"},
@@ -593,6 +596,10 @@ func maxOutputTokensLimit(model string) int {
 	return maxOutputTokensUpperBound
 }
 
+func isAntigravityOpus46Model(model string) bool {
+	return strings.HasPrefix(strings.ToLower(model), "claude-opus-4-6")
+}
+
 func buildGenerationConfig(req *ClaudeRequest) *GeminiGenerationConfig {
 	maxLimit := maxOutputTokensLimit(req.Model)
 	config := &GeminiGenerationConfig{
@@ -606,25 +613,36 @@ func buildGenerationConfig(req *ClaudeRequest) *GeminiGenerationConfig {
 	}
 
 	// Thinking 配置
-	if req.Thinking != nil && req.Thinking.Type == "enabled" {
+	if req.Thinking != nil && (req.Thinking.Type == "enabled" || req.Thinking.Type == "adaptive") {
 		config.ThinkingConfig = &GeminiThinkingConfig{
 			IncludeThoughts: true,
 		}
+
+		// - thinking.type=enabled：budget_tokens>0 用显式预算
+		// - thinking.type=adaptive：仅在 Antigravity 的 Opus 4.6 上覆写为 （24576）
+		budget := -1
 		if req.Thinking.BudgetTokens > 0 {
-			budget := req.Thinking.BudgetTokens
+			budget = req.Thinking.BudgetTokens
+		}
+		if req.Thinking.Type == "adaptive" && isAntigravityOpus46Model(req.Model) {
+			budget = ClaudeAdaptiveHighThinkingBudgetTokens
+		}
+
+		// 正预算需要做上限与 max_tokens 约束；动态预算（-1）直接透传给上游。
+		if budget > 0 {
 			// gemini-2.5-flash 上限
 			if strings.Contains(req.Model, "gemini-2.5-flash") && budget > Gemini25FlashThinkingBudgetLimit {
 				budget = Gemini25FlashThinkingBudgetLimit
 			}
-			config.ThinkingConfig.ThinkingBudget = budget
 
-			// 自动修正：max_tokens 必须大于 budget_tokens
+			// 自动修正：max_tokens 必须大于 budget_tokens（Claude 上游要求）
 			if adjusted, ok := ensureMaxTokensGreaterThanBudget(config.MaxOutputTokens, budget); ok {
 				log.Printf("[Antigravity] Auto-adjusted max_tokens from %d to %d (must be > budget_tokens=%d)",
 					config.MaxOutputTokens, adjusted, budget)
 				config.MaxOutputTokens = adjusted
 			}
 		}
+		config.ThinkingConfig.ThinkingBudget = budget
 	}
 
 	if config.MaxOutputTokens > maxLimit {

backend/internal/pkg/antigravity/request_transformer_test.go

@@ -259,3 +259,93 @@ func TestBuildTools_CustomTypeTools(t *testing.T) {
 		})
 	}
 }
+
+func TestBuildGenerationConfig_ThinkingDynamicBudget(t *testing.T) {
+	tests := []struct {
+		name        string
+		model       string
+		thinking    *ThinkingConfig
+		wantBudget  int
+		wantPresent bool
+	}{
+		{
+			name:        "enabled without budget defaults to dynamic (-1)",
+			model:       "claude-opus-4-6-thinking",
+			thinking:    &ThinkingConfig{Type: "enabled"},
+			wantBudget:  -1,
+			wantPresent: true,
+		},
+		{
+			name:        "enabled with budget uses the provided value",
+			model:       "claude-opus-4-6-thinking",
+			thinking:    &ThinkingConfig{Type: "enabled", BudgetTokens: 1024},
+			wantBudget:  1024,
+			wantPresent: true,
+		},
+		{
+			name:        "enabled with -1 budget uses dynamic (-1)",
+			model:       "claude-opus-4-6-thinking",
+			thinking:    &ThinkingConfig{Type: "enabled", BudgetTokens: -1},
+			wantBudget:  -1,
+			wantPresent: true,
+		},
+		{
+			name:        "adaptive on opus4.6 maps to high budget (24576)",
+			model:       "claude-opus-4-6-thinking",
+			thinking:    &ThinkingConfig{Type: "adaptive", BudgetTokens: 20000},
+			wantBudget:  ClaudeAdaptiveHighThinkingBudgetTokens,
+			wantPresent: true,
+		},
+		{
+			name:        "adaptive on non-opus model keeps default dynamic (-1)",
+			model:       "claude-sonnet-4-5-thinking",
+			thinking:    &ThinkingConfig{Type: "adaptive"},
+			wantBudget:  -1,
+			wantPresent: true,
+		},
+		{
+			name:        "disabled does not emit thinkingConfig",
+			model:       "claude-opus-4-6-thinking",
+			thinking:    &ThinkingConfig{Type: "disabled", BudgetTokens: 1024},
+			wantBudget:  0,
+			wantPresent: false,
+		},
+		{
+			name:        "nil thinking does not emit thinkingConfig",
+			model:       "claude-opus-4-6-thinking",
+			thinking:    nil,
+			wantBudget:  0,
+			wantPresent: false,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			req := &ClaudeRequest{
+				Model:    tt.model,
+				Thinking: tt.thinking,
+			}
+			cfg := buildGenerationConfig(req)
+			if cfg == nil {
+				t.Fatalf("expected non-nil generationConfig")
+			}
+
+			if tt.wantPresent {
+				if cfg.ThinkingConfig == nil {
+					t.Fatalf("expected thinkingConfig to be present")
+				}
+				if !cfg.ThinkingConfig.IncludeThoughts {
+					t.Fatalf("expected includeThoughts=true")
+				}
+				if cfg.ThinkingConfig.ThinkingBudget != tt.wantBudget {
+					t.Fatalf("expected thinkingBudget=%d, got %d", tt.wantBudget, cfg.ThinkingConfig.ThinkingBudget)
+				}
+				return
+			}
+
+			if cfg.ThinkingConfig != nil {
+				t.Fatalf("expected thinkingConfig to be nil, got %+v", cfg.ThinkingConfig)
+			}
+		})
+	}
+}

backend/internal/pkg/antigravity/response_transformer.go

@@ -282,7 +282,7 @@ func (p *NonStreamingProcessor) buildResponse(geminiResp *GeminiResponse, respon
 	if geminiResp.UsageMetadata != nil {
 		cached := geminiResp.UsageMetadata.CachedContentTokenCount
 		usage.InputTokens = geminiResp.UsageMetadata.PromptTokenCount - cached
-		usage.OutputTokens = geminiResp.UsageMetadata.CandidatesTokenCount
+		usage.OutputTokens = geminiResp.UsageMetadata.CandidatesTokenCount + geminiResp.UsageMetadata.ThoughtsTokenCount
 		usage.CacheReadInputTokens = cached
 	}
 

backend/internal/pkg/antigravity/stream_transformer.go

@@ -85,7 +85,7 @@ func (p *StreamingProcessor) ProcessLine(line string) []byte {
 	if geminiResp.UsageMetadata != nil {
 		cached := geminiResp.UsageMetadata.CachedContentTokenCount
 		p.inputTokens = geminiResp.UsageMetadata.PromptTokenCount - cached
-		p.outputTokens = geminiResp.UsageMetadata.CandidatesTokenCount
+		p.outputTokens = geminiResp.UsageMetadata.CandidatesTokenCount + geminiResp.UsageMetadata.ThoughtsTokenCount
 		p.cacheReadTokens = cached
 	}
 
@@ -146,7 +146,7 @@ func (p *StreamingProcessor) emitMessageStart(v1Resp *V1InternalResponse) []byte
 	if v1Resp.Response.UsageMetadata != nil {
 		cached := v1Resp.Response.UsageMetadata.CachedContentTokenCount
 		usage.InputTokens = v1Resp.Response.UsageMetadata.PromptTokenCount - cached
-		usage.OutputTokens = v1Resp.Response.UsageMetadata.CandidatesTokenCount
+		usage.OutputTokens = v1Resp.Response.UsageMetadata.CandidatesTokenCount + v1Resp.Response.UsageMetadata.ThoughtsTokenCount
 		usage.CacheReadInputTokens = cached
 	}
 

backend/internal/pkg/logger/slog_handler_test.go

@@ -15,7 +15,6 @@ type captureState struct {
 }
 
 type capturedWrite struct {
-	entry  zapcore.Entry
 	fields []zapcore.Field
 }
 
@@ -51,7 +50,6 @@ func (c *captureCore) Write(entry zapcore.Entry, fields []zapcore.Field) error {
 	allFields = append(allFields, c.withFields...)
 	allFields = append(allFields, fields...)
 	c.state.writes = append(c.state.writes, capturedWrite{
-		entry:  entry,
 		fields: allFields,
 	})
 	return nil

backend/internal/repository/account_repo.go

@@ -448,7 +448,12 @@ func (r *accountRepository) ListWithFilters(ctx context.Context, params paginati
 		q = q.Where(dbaccount.TypeEQ(accountType))
 	}
 	if status != "" {
-		q = q.Where(dbaccount.StatusEQ(status))
+		switch status {
+		case "rate_limited":
+			q = q.Where(dbaccount.RateLimitResetAtGT(time.Now()))
+		default:
+			q = q.Where(dbaccount.StatusEQ(status))
+		}
 	}
 	if search != "" {
 		q = q.Where(dbaccount.NameContainsFold(search))

backend/internal/repository/error_passthrough_repo.go

@@ -54,7 +54,8 @@ func (r *errorPassthroughRepository) Create(ctx context.Context, rule *model.Err
 		SetPriority(rule.Priority).
 		SetMatchMode(rule.MatchMode).
 		SetPassthroughCode(rule.PassthroughCode).
-		SetPassthroughBody(rule.PassthroughBody)
+		SetPassthroughBody(rule.PassthroughBody).
+		SetSkipMonitoring(rule.SkipMonitoring)
 
 	if len(rule.ErrorCodes) > 0 {
 		builder.SetErrorCodes(rule.ErrorCodes)
@@ -90,7 +91,8 @@ func (r *errorPassthroughRepository) Update(ctx context.Context, rule *model.Err
 		SetPriority(rule.Priority).
 		SetMatchMode(rule.MatchMode).
 		SetPassthroughCode(rule.PassthroughCode).
-		SetPassthroughBody(rule.PassthroughBody)
+		SetPassthroughBody(rule.PassthroughBody).
+		SetSkipMonitoring(rule.SkipMonitoring)
 
 	// 处理可选字段
 	if len(rule.ErrorCodes) > 0 {
@@ -149,6 +151,7 @@ func (r *errorPassthroughRepository) toModel(e *ent.ErrorPassthroughRule) *model
 		Platforms:       e.Platforms,
 		PassthroughCode: e.PassthroughCode,
 		PassthroughBody: e.PassthroughBody,
+		SkipMonitoring:  e.SkipMonitoring,
 		CreatedAt:       e.CreatedAt,
 		UpdatedAt:       e.UpdatedAt,
 	}

backend/internal/repository/redeem_code_repo.go

@@ -6,6 +6,7 @@ import (
 
 	dbent "github.com/Wei-Shaw/sub2api/ent"
 	"github.com/Wei-Shaw/sub2api/ent/redeemcode"
+	"github.com/Wei-Shaw/sub2api/ent/user"
 	"github.com/Wei-Shaw/sub2api/internal/pkg/pagination"
 	"github.com/Wei-Shaw/sub2api/internal/service"
 )
@@ -106,7 +107,12 @@ func (r *redeemCodeRepository) ListWithFilters(ctx context.Context, params pagin
 		q = q.Where(redeemcode.StatusEQ(status))
 	}
 	if search != "" {
-		q = q.Where(redeemcode.CodeContainsFold(search))
+		q = q.Where(
+			redeemcode.Or(
+				redeemcode.CodeContainsFold(search),
+				redeemcode.HasUserWith(user.EmailContainsFold(search)),
+			),
+		)
 	}
 
 	total, err := q.Count(ctx)

backend/internal/repository/user_repo.go

@@ -10,6 +10,7 @@ import (
 	"time"
 
 	dbent "github.com/Wei-Shaw/sub2api/ent"
+	"github.com/Wei-Shaw/sub2api/ent/apikey"
 	dbuser "github.com/Wei-Shaw/sub2api/ent/user"
 	"github.com/Wei-Shaw/sub2api/ent/userallowedgroup"
 	"github.com/Wei-Shaw/sub2api/ent/usersubscription"
@@ -191,6 +192,7 @@ func (r *userRepository) ListWithFilters(ctx context.Context, params pagination.
 				dbuser.EmailContainsFold(filters.Search),
 				dbuser.UsernameContainsFold(filters.Search),
 				dbuser.NotesContainsFold(filters.Search),
+				dbuser.HasAPIKeysWith(apikey.KeyContainsFold(filters.Search)),
 			),
 		)
 	}

backend/internal/server/routes/admin.go

@@ -290,6 +290,7 @@ func registerAntigravityOAuthRoutes(admin *gin.RouterGroup, h *handler.Handlers)
 	{
 		antigravity.POST("/oauth/auth-url", h.Admin.AntigravityOAuth.GenerateAuthURL)
 		antigravity.POST("/oauth/exchange-code", h.Admin.AntigravityOAuth.ExchangeCode)
+		antigravity.POST("/oauth/refresh-token", h.Admin.AntigravityOAuth.RefreshToken)
 	}
 }
 

backend/internal/service/antigravity_gateway_service.go

@@ -8,6 +8,7 @@ import (
 	"errors"
 	"fmt"
 	"io"
+	"log"
 	"log/slog"
 	mathrand "math/rand"
 	"net"
@@ -15,6 +16,7 @@ import (
 	"os"
 	"strconv"
 	"strings"
+	"sync"
 	"sync/atomic"
 	"time"
 
@@ -41,6 +43,12 @@ const (
 	antigravitySmartRetryMaxAttempts    = 1                // 智能重试最大次数（仅重试 1 次，防止重复限流/长期等待）
 	antigravityDefaultRateLimitDuration = 30 * time.Second // 默认限流时间（无 retryDelay 时使用）
 
+	// MODEL_CAPACITY_EXHAUSTED 专用重试参数
+	// 模型容量不足时，所有账号共享同一容量池，切换账号无意义
+	// 使用固定 1s 间隔重试，最多重试 60 次
+	antigravityModelCapacityRetryMaxAttempts = 60
+	antigravityModelCapacityRetryWait        = 1 * time.Second
+
 	// Google RPC 状态和类型常量
 	googleRPCStatusResourceExhausted      = "RESOURCE_EXHAUSTED"
 	googleRPCStatusUnavailable            = "UNAVAILABLE"
@@ -61,6 +69,9 @@ const (
 	// 单账号 503 退避重试：原地重试的总累计等待时间上限
 	// 超过此上限将不再重试，直接返回 503
 	antigravitySingleAccountSmartRetryTotalMaxWait = 30 * time.Second
+
+	// MODEL_CAPACITY_EXHAUSTED 全局去重：重试全部失败后的 cooldown 时间
+	antigravityModelCapacityCooldown = 10 * time.Second
 )
 
 // antigravityPassthroughErrorMessages 透传给客户端的错误消息白名单（小写）
@@ -69,8 +80,15 @@ var antigravityPassthroughErrorMessages = []string{
 	"prompt is too long",
 }
 
+// MODEL_CAPACITY_EXHAUSTED 全局去重：避免多个并发请求同时对同一模型进行容量耗尽重试
+var (
+	modelCapacityExhaustedMu    sync.RWMutex
+	modelCapacityExhaustedUntil = make(map[string]time.Time) // modelName -> cooldown until
+)
+
 const (
 	antigravityBillingModelEnv    = "GATEWAY_ANTIGRAVITY_BILL_WITH_MAPPED_MODEL"
+	antigravityForwardBaseURLEnv  = "GATEWAY_ANTIGRAVITY_FORWARD_BASE_URL"
 	antigravityFallbackSecondsEnv = "GATEWAY_ANTIGRAVITY_FALLBACK_COOLDOWN_SECONDS"
 )
 
@@ -132,6 +150,20 @@ type antigravityRetryLoopResult struct {
 	resp *http.Response
 }
 
+// resolveAntigravityForwardBaseURL 解析转发用 base URL。
+// 默认使用 daily（ForwardBaseURLs 的首个地址）；当环境变量为 prod 时使用第二个地址。
+func resolveAntigravityForwardBaseURL() string {
+	baseURLs := antigravity.ForwardBaseURLs()
+	if len(baseURLs) == 0 {
+		return ""
+	}
+	mode := strings.ToLower(strings.TrimSpace(os.Getenv(antigravityForwardBaseURLEnv)))
+	if mode == "prod" && len(baseURLs) > 1 {
+		return baseURLs[1]
+	}
+	return baseURLs[0]
+}
+
 // smartRetryAction 智能重试的处理结果
 type smartRetryAction int
 
@@ -159,7 +191,7 @@ func (s *AntigravityGatewayService) handleSmartRetry(p antigravityRetryLoopParam
 	}
 
 	// 判断是否触发智能重试
-	shouldSmartRetry, shouldRateLimitModel, waitDuration, modelName := shouldTriggerAntigravitySmartRetry(p.account, respBody)
+	shouldSmartRetry, shouldRateLimitModel, waitDuration, modelName, isModelCapacityExhausted := shouldTriggerAntigravitySmartRetry(p.account, respBody)
 
 	// 情况1: retryDelay >= 阈值，限流模型并切换账号
 	if shouldRateLimitModel {
@@ -196,20 +228,48 @@ func (s *AntigravityGatewayService) handleSmartRetry(p antigravityRetryLoopParam
 		}
 	}
 
-	// 情况2: retryDelay < 阈值，智能重试（最多 antigravitySmartRetryMaxAttempts 次）
+	// 情况2: retryDelay < 阈值（或 MODEL_CAPACITY_EXHAUSTED），智能重试
 	if shouldSmartRetry {
 		var lastRetryResp *http.Response
 		var lastRetryBody []byte
 
-		for attempt := 1; attempt <= antigravitySmartRetryMaxAttempts; attempt++ {
-			logger.LegacyPrintf("service.antigravity_gateway", "%s status=%d oauth_smart_retry attempt=%d/%d delay=%v model=%s account=%d",
-				p.prefix, resp.StatusCode, attempt, antigravitySmartRetryMaxAttempts, waitDuration, modelName, p.account.ID)
+		// MODEL_CAPACITY_EXHAUSTED 使用独立的重试参数（60 次，固定 1s 间隔）
+		maxAttempts := antigravitySmartRetryMaxAttempts
+		if isModelCapacityExhausted {
+			maxAttempts = antigravityModelCapacityRetryMaxAttempts
+			waitDuration = antigravityModelCapacityRetryWait
+
+			// 全局去重：如果其他 goroutine 已在重试同一模型且尚在 cooldown 中，直接返回 503
+			if modelName != "" {
+				modelCapacityExhaustedMu.RLock()
+				cooldownUntil, exists := modelCapacityExhaustedUntil[modelName]
+				modelCapacityExhaustedMu.RUnlock()
+				if exists && time.Now().Before(cooldownUntil) {
+					log.Printf("%s status=%d model_capacity_exhausted_dedup model=%s account=%d cooldown_until=%v (skip retry)",
+						p.prefix, resp.StatusCode, modelName, p.account.ID, cooldownUntil.Format("15:04:05"))
+					return &smartRetryResult{
+						action: smartRetryActionBreakWithResp,
+						resp: &http.Response{
+							StatusCode: resp.StatusCode,
+							Header:     resp.Header.Clone(),
+							Body:       io.NopCloser(bytes.NewReader(respBody)),
+						},
+					}
+				}
+			}
+		}
+
+		for attempt := 1; attempt <= maxAttempts; attempt++ {
+			log.Printf("%s status=%d oauth_smart_retry attempt=%d/%d delay=%v model=%s account=%d",
+				p.prefix, resp.StatusCode, attempt, maxAttempts, waitDuration, modelName, p.account.ID)
 
+			timer := time.NewTimer(waitDuration)
 			select {
 			case <-p.ctx.Done():
-				logger.LegacyPrintf("service.antigravity_gateway", "%s status=context_canceled_during_smart_retry", p.prefix)
+				timer.Stop()
+				log.Printf("%s status=context_canceled_during_smart_retry", p.prefix)
 				return &smartRetryResult{action: smartRetryActionBreakWithResp, err: p.ctx.Err()}
-			case <-time.After(waitDuration):
+			case <-timer.C:
 			}
 
 			// 智能重试：创建新请求
@@ -229,13 +289,19 @@ func (s *AntigravityGatewayService) handleSmartRetry(p antigravityRetryLoopParam
 
 			retryResp, retryErr := p.httpUpstream.Do(retryReq, p.proxyURL, p.account.ID, p.account.Concurrency)
 			if retryErr == nil && retryResp != nil && retryResp.StatusCode != http.StatusTooManyRequests && retryResp.StatusCode != http.StatusServiceUnavailable {
-				logger.LegacyPrintf("service.antigravity_gateway", "%s status=%d smart_retry_success attempt=%d/%d", p.prefix, retryResp.StatusCode, attempt, antigravitySmartRetryMaxAttempts)
+				log.Printf("%s status=%d smart_retry_success attempt=%d/%d", p.prefix, retryResp.StatusCode, attempt, maxAttempts)
+				// 重试成功，清除 MODEL_CAPACITY_EXHAUSTED cooldown
+				if isModelCapacityExhausted && modelName != "" {
+					modelCapacityExhaustedMu.Lock()
+					delete(modelCapacityExhaustedUntil, modelName)
+					modelCapacityExhaustedMu.Unlock()
+				}
 				return &smartRetryResult{action: smartRetryActionBreakWithResp, resp: retryResp}
 			}
 
 			// 网络错误时，继续重试
 			if retryErr != nil || retryResp == nil {
-				logger.LegacyPrintf("service.antigravity_gateway", "%s status=smart_retry_network_error attempt=%d/%d error=%v", p.prefix, attempt, antigravitySmartRetryMaxAttempts, retryErr)
+				log.Printf("%s status=smart_retry_network_error attempt=%d/%d error=%v", p.prefix, attempt, maxAttempts, retryErr)
 				continue
 			}
 
@@ -245,13 +311,13 @@ func (s *AntigravityGatewayService) handleSmartRetry(p antigravityRetryLoopParam
 			}
 			lastRetryResp = retryResp
 			if retryResp != nil {
-				lastRetryBody, _ = io.ReadAll(io.LimitReader(retryResp.Body, 2<<20))
+				lastRetryBody, _ = io.ReadAll(io.LimitReader(retryResp.Body, 8<<10))
 				_ = retryResp.Body.Close()
 			}
 
-			// 解析新的重试信息，用于下次重试的等待时间
-			if attempt < antigravitySmartRetryMaxAttempts && lastRetryBody != nil {
-				newShouldRetry, _, newWaitDuration, _ := shouldTriggerAntigravitySmartRetry(p.account, lastRetryBody)
+			// 解析新的重试信息，用于下次重试的等待时间（MODEL_CAPACITY_EXHAUSTED 使用固定循环，跳过）
+			if !isModelCapacityExhausted && attempt < maxAttempts && lastRetryBody != nil {
+				newShouldRetry, _, newWaitDuration, _, _ := shouldTriggerAntigravitySmartRetry(p.account, lastRetryBody)
 				if newShouldRetry && newWaitDuration > 0 {
 					waitDuration = newWaitDuration
 				}
@@ -268,6 +334,27 @@ func (s *AntigravityGatewayService) handleSmartRetry(p antigravityRetryLoopParam
 			retryBody = respBody
 		}
 
+		// MODEL_CAPACITY_EXHAUSTED：模型容量不足，切换账号无意义
+		// 直接返回上游错误响应，不设置模型限流，不切换账号
+		if isModelCapacityExhausted {
+			// 设置 cooldown，让后续请求快速失败，避免重复重试
+			if modelName != "" {
+				modelCapacityExhaustedMu.Lock()
+				modelCapacityExhaustedUntil[modelName] = time.Now().Add(antigravityModelCapacityCooldown)
+				modelCapacityExhaustedMu.Unlock()
+			}
+			log.Printf("%s status=%d smart_retry_exhausted_model_capacity attempts=%d model=%s account=%d body=%s (model capacity exhausted, not switching account)",
+				p.prefix, resp.StatusCode, maxAttempts, modelName, p.account.ID, truncateForLog(retryBody, 200))
+			return &smartRetryResult{
+				action: smartRetryActionBreakWithResp,
+				resp: &http.Response{
+					StatusCode: resp.StatusCode,
+					Header:     resp.Header.Clone(),
+					Body:       io.NopCloser(bytes.NewReader(retryBody)),
+				},
+			}
+		}
+
 		// 单账号 503 退避重试模式：智能重试耗尽后不设限流、不切换账号，
 		// 直接返回 503 让 Handler 层的单账号退避循环做最终处理。
 		if resp.StatusCode == http.StatusServiceUnavailable && isSingleAccountRetry(p.ctx) {
@@ -283,8 +370,8 @@ func (s *AntigravityGatewayService) handleSmartRetry(p antigravityRetryLoopParam
 			}
 		}
 
-		logger.LegacyPrintf("service.antigravity_gateway", "%s status=%d smart_retry_exhausted attempts=%d model=%s account=%d upstream_retry_delay=%v body=%s (switch account)",
-			p.prefix, resp.StatusCode, antigravitySmartRetryMaxAttempts, modelName, p.account.ID, rateLimitDuration, truncateForLog(retryBody, 200))
+		log.Printf("%s status=%d smart_retry_exhausted attempts=%d model=%s account=%d upstream_retry_delay=%v body=%s (switch account)",
+			p.prefix, resp.StatusCode, maxAttempts, modelName, p.account.ID, rateLimitDuration, truncateForLog(retryBody, 200))
 
 		resetAt := time.Now().Add(rateLimitDuration)
 		if p.accountRepo != nil && modelName != "" {
@@ -368,11 +455,13 @@ func (s *AntigravityGatewayService) handleSingleAccountRetryInPlace(
 		logger.LegacyPrintf("service.antigravity_gateway", "%s status=%d single_account_503_retry attempt=%d/%d delay=%v total_waited=%v model=%s account=%d",
 			p.prefix, resp.StatusCode, attempt, antigravitySingleAccountSmartRetryMaxAttempts, waitDuration, totalWaited, modelName, p.account.ID)
 
+		timer := time.NewTimer(waitDuration)
 		select {
 		case <-p.ctx.Done():
+			timer.Stop()
 			logger.LegacyPrintf("service.antigravity_gateway", "%s status=context_canceled_during_single_account_retry", p.prefix)
 			return &smartRetryResult{action: smartRetryActionBreakWithResp, err: p.ctx.Err()}
-		case <-time.After(waitDuration):
+		case <-timer.C:
 		}
 		totalWaited += waitDuration
 
@@ -406,12 +495,12 @@ func (s *AntigravityGatewayService) handleSingleAccountRetryInPlace(
 			_ = lastRetryResp.Body.Close()
 		}
 		lastRetryResp = retryResp
-		lastRetryBody, _ = io.ReadAll(io.LimitReader(retryResp.Body, 2<<20))
+		lastRetryBody, _ = io.ReadAll(io.LimitReader(retryResp.Body, 8<<10))
 		_ = retryResp.Body.Close()
 
 		// 解析新的重试信息，更新下次等待时间
 		if attempt < antigravitySingleAccountSmartRetryMaxAttempts && lastRetryBody != nil {
-			_, _, newWaitDuration, _ := shouldTriggerAntigravitySmartRetry(p.account, lastRetryBody)
+			_, _, newWaitDuration, _, _ := shouldTriggerAntigravitySmartRetry(p.account, lastRetryBody)
 			if newWaitDuration > 0 {
 				waitDuration = newWaitDuration
 				if waitDuration > antigravitySingleAccountSmartRetryMaxWait {
@@ -467,10 +556,11 @@ func (s *AntigravityGatewayService) antigravityRetryLoop(p antigravityRetryLoopP
 		}
 	}
 
-	availableURLs := antigravity.DefaultURLAvailability.GetAvailableURLs()
-	if len(availableURLs) == 0 {
-		availableURLs = antigravity.BaseURLs
+	baseURL := resolveAntigravityForwardBaseURL()
+	if baseURL == "" {
+		return nil, errors.New("no antigravity forward base url configured")
 	}
+	availableURLs := []string{baseURL}
 
 	var resp *http.Response
 	var usedBaseURL string
@@ -908,11 +998,11 @@ func (s *AntigravityGatewayService) TestConnection(ctx context.Context, account
 		proxyURL = account.Proxy.URL()
 	}
 
-	// URL fallback 循环
-	availableURLs := antigravity.DefaultURLAvailability.GetAvailableURLs()
-	if len(availableURLs) == 0 {
-		availableURLs = antigravity.BaseURLs // 所有 URL 都不可用时，重试所有
+	baseURL := resolveAntigravityForwardBaseURL()
+	if baseURL == "" {
+		return nil, errors.New("no antigravity forward base url configured")
 	}
+	availableURLs := []string{baseURL}
 
 	var lastErr error
 	for urlIdx, baseURL := range availableURLs {
@@ -1217,7 +1307,7 @@ func (s *AntigravityGatewayService) Forward(ctx context.Context, c *gin.Context,
 		return nil, s.writeClaudeError(c, http.StatusForbidden, "permission_error", fmt.Sprintf("model %s not in whitelist", claudeReq.Model))
 	}
 	// 应用 thinking 模式自动后缀：如果 thinking 开启且目标是 claude-sonnet-4-5，自动改为 thinking 版本
-	thinkingEnabled := claudeReq.Thinking != nil && claudeReq.Thinking.Type == "enabled"
+	thinkingEnabled := claudeReq.Thinking != nil && (claudeReq.Thinking.Type == "enabled" || claudeReq.Thinking.Type == "adaptive")
 	mappedModel = applyThinkingModelSuffix(mappedModel, thinkingEnabled)
 
 	// 获取 access_token
@@ -1373,7 +1463,7 @@ func (s *AntigravityGatewayService) Forward(ctx context.Context, c *gin.Context,
 					break
 				}
 
-				retryBody, _ := io.ReadAll(io.LimitReader(retryResp.Body, 2<<20))
+				retryBody, _ := io.ReadAll(io.LimitReader(retryResp.Body, 8<<10))
 				_ = retryResp.Body.Close()
 				if retryResp.StatusCode == http.StatusTooManyRequests {
 					retryBaseURL := ""
@@ -1454,6 +1544,27 @@ func (s *AntigravityGatewayService) Forward(ctx context.Context, c *gin.Context,
 
 			s.handleUpstreamError(ctx, prefix, account, resp.StatusCode, resp.Header, respBody, originalModel, 0, "", isStickySession)
 
+			// 精确匹配服务端配置类 400 错误，触发同账号重试 + failover
+			if resp.StatusCode == http.StatusBadRequest {
+				msg := strings.ToLower(strings.TrimSpace(extractAntigravityErrorMessage(respBody)))
+				if isGoogleProjectConfigError(msg) {
+					upstreamMsg := sanitizeUpstreamErrorMessage(strings.TrimSpace(extractAntigravityErrorMessage(respBody)))
+					upstreamDetail := s.getUpstreamErrorDetail(respBody)
+					log.Printf("%s status=400 google_config_error failover=true upstream_message=%q account=%d", prefix, upstreamMsg, account.ID)
+					appendOpsUpstreamError(c, OpsUpstreamErrorEvent{
+						Platform:           account.Platform,
+						AccountID:          account.ID,
+						AccountName:        account.Name,
+						UpstreamStatusCode: resp.StatusCode,
+						UpstreamRequestID:  resp.Header.Get("x-request-id"),
+						Kind:               "failover",
+						Message:            upstreamMsg,
+						Detail:             upstreamDetail,
+					})
+					return nil, &UpstreamFailoverError{StatusCode: resp.StatusCode, ResponseBody: respBody, RetryableOnSameAccount: true}
+				}
+			}
+
 			if s.shouldFailoverUpstreamError(resp.StatusCode) {
 				upstreamMsg := strings.TrimSpace(extractAntigravityErrorMessage(respBody))
 				upstreamMsg = sanitizeUpstreamErrorMessage(upstreamMsg)
@@ -1994,6 +2105,22 @@ func (s *AntigravityGatewayService) ForwardGemini(ctx context.Context, c *gin.Co
 		// Always record upstream context for Ops error logs, even when we will failover.
 		setOpsUpstreamError(c, resp.StatusCode, upstreamMsg, upstreamDetail)
 
+		// 精确匹配服务端配置类 400 错误，触发同账号重试 + failover
+		if resp.StatusCode == http.StatusBadRequest && isGoogleProjectConfigError(strings.ToLower(upstreamMsg)) {
+			log.Printf("%s status=400 google_config_error failover=true upstream_message=%q account=%d", prefix, upstreamMsg, account.ID)
+			appendOpsUpstreamError(c, OpsUpstreamErrorEvent{
+				Platform:           account.Platform,
+				AccountID:          account.ID,
+				AccountName:        account.Name,
+				UpstreamStatusCode: resp.StatusCode,
+				UpstreamRequestID:  requestID,
+				Kind:               "failover",
+				Message:            upstreamMsg,
+				Detail:             upstreamDetail,
+			})
+			return nil, &UpstreamFailoverError{StatusCode: resp.StatusCode, ResponseBody: unwrappedForOps, RetryableOnSameAccount: true}
+		}
+
 		if s.shouldFailoverUpstreamError(resp.StatusCode) {
 			appendOpsUpstreamError(c, OpsUpstreamErrorEvent{
 				Platform:           account.Platform,
@@ -2089,6 +2216,44 @@ func (s *AntigravityGatewayService) shouldFailoverUpstreamError(statusCode int)
 	}
 }
 
+// isGoogleProjectConfigError 判断（已提取的小写）错误消息是否属于 Google 服务端配置类问题。
+// 只精确匹配已知的服务端侧错误，避免对客户端请求错误做无意义重试。
+// 适用于所有走 Google 后端的平台（Antigravity、Gemini）。
+func isGoogleProjectConfigError(lowerMsg string) bool {
+	// Google 间歇性 Bug：Project ID 有效但被临时识别失败
+	return strings.Contains(lowerMsg, "invalid project resource name")
+}
+
+// googleConfigErrorCooldown 服务端配置类 400 错误的临时封禁时长
+const googleConfigErrorCooldown = 1 * time.Minute
+
+// tempUnscheduleGoogleConfigError 对服务端配置类 400 错误触发临时封禁，
+// 避免短时间内反复调度到同一个有问题的账号。
+func tempUnscheduleGoogleConfigError(ctx context.Context, repo AccountRepository, accountID int64, logPrefix string) {
+	until := time.Now().Add(googleConfigErrorCooldown)
+	reason := "400: invalid project resource name (auto temp-unschedule 1m)"
+	if err := repo.SetTempUnschedulable(ctx, accountID, until, reason); err != nil {
+		log.Printf("%s temp_unschedule_failed account=%d error=%v", logPrefix, accountID, err)
+	} else {
+		log.Printf("%s temp_unscheduled account=%d until=%v reason=%q", logPrefix, accountID, until.Format("15:04:05"), reason)
+	}
+}
+
+// emptyResponseCooldown 空流式响应的临时封禁时长
+const emptyResponseCooldown = 1 * time.Minute
+
+// tempUnscheduleEmptyResponse 对空流式响应触发临时封禁，
+// 避免短时间内反复调度到同一个返回空响应的账号。
+func tempUnscheduleEmptyResponse(ctx context.Context, repo AccountRepository, accountID int64, logPrefix string) {
+	until := time.Now().Add(emptyResponseCooldown)
+	reason := "empty stream response (auto temp-unschedule 1m)"
+	if err := repo.SetTempUnschedulable(ctx, accountID, until, reason); err != nil {
+		log.Printf("%s temp_unschedule_failed account=%d error=%v", logPrefix, accountID, err)
+	} else {
+		log.Printf("%s temp_unscheduled account=%d until=%v reason=%q", logPrefix, accountID, until.Format("15:04:05"), reason)
+	}
+}
+
 // sleepAntigravityBackoffWithContext 带 context 取消检查的退避等待
 // 返回 true 表示正常完成等待，false 表示 context 已取消
 func sleepAntigravityBackoffWithContext(ctx context.Context, attempt int) bool {
@@ -2105,10 +2270,12 @@ func sleepAntigravityBackoffWithContext(ctx context.Context, attempt int) bool {
 		sleepFor = 0
 	}
 
+	timer := time.NewTimer(sleepFor)
 	select {
 	case <-ctx.Done():
+		timer.Stop()
 		return false
-	case <-time.After(sleepFor):
+	case <-timer.C:
 		return true
 	}
 }
@@ -2153,8 +2320,9 @@ func antigravityFallbackCooldownSeconds() (time.Duration, bool) {
 
 // antigravitySmartRetryInfo 智能重试所需的信息
 type antigravitySmartRetryInfo struct {
-	RetryDelay time.Duration // 重试延迟时间
-	ModelName  string        // 限流的模型名称（如 "claude-sonnet-4-5"）
+	RetryDelay               time.Duration // 重试延迟时间
+	ModelName                string        // 限流的模型名称（如 "claude-sonnet-4-5"）
+	IsModelCapacityExhausted bool          // 是否为模型容量不足（MODEL_CAPACITY_EXHAUSTED）
 }
 
 // parseAntigravitySmartRetryInfo 解析 Google RPC RetryInfo 和 ErrorInfo 信息
@@ -2269,31 +2437,40 @@ func parseAntigravitySmartRetryInfo(body []byte) *antigravitySmartRetryInfo {
 	}
 
 	return &antigravitySmartRetryInfo{
-		RetryDelay: retryDelay,
-		ModelName:  modelName,
+		RetryDelay:               retryDelay,
+		ModelName:                modelName,
+		IsModelCapacityExhausted: hasModelCapacityExhausted,
 	}
 }
 
 // shouldTriggerAntigravitySmartRetry 判断是否应该触发智能重试
 // 返回：
-//   - shouldRetry: 是否应该智能重试（retryDelay < antigravityRateLimitThreshold）
-//   - shouldRateLimitModel: 是否应该限流模型（retryDelay >= antigravityRateLimitThreshold）
-//   - waitDuration: 等待时间（智能重试时使用，shouldRateLimitModel=true 时为 0）
+//   - shouldRetry: 是否应该智能重试（retryDelay < antigravityRateLimitThreshold，或 MODEL_CAPACITY_EXHAUSTED）
+//   - shouldRateLimitModel: 是否应该限流模型并切换账号（仅 RATE_LIMIT_EXCEEDED 且 retryDelay >= 阈值）
+//   - waitDuration: 等待时间
 //   - modelName: 限流的模型名称
-func shouldTriggerAntigravitySmartRetry(account *Account, respBody []byte) (shouldRetry bool, shouldRateLimitModel bool, waitDuration time.Duration, modelName string) {
+//   - isModelCapacityExhausted: 是否为模型容量不足（MODEL_CAPACITY_EXHAUSTED）
+func shouldTriggerAntigravitySmartRetry(account *Account, respBody []byte) (shouldRetry bool, shouldRateLimitModel bool, waitDuration time.Duration, modelName string, isModelCapacityExhausted bool) {
 	if account.Platform != PlatformAntigravity {
-		return false, false, 0, ""
+		return false, false, 0, "", false
 	}
 
 	info := parseAntigravitySmartRetryInfo(respBody)
 	if info == nil {
-		return false, false, 0, ""
+		return false, false, 0, "", false
 	}
 
+	// MODEL_CAPACITY_EXHAUSTED（模型容量不足）：所有账号共享同一模型容量池
+	// 切换账号无意义，使用固定 1s 间隔重试
+	if info.IsModelCapacityExhausted {
+		return true, false, antigravityModelCapacityRetryWait, info.ModelName, true
+	}
+
+	// RATE_LIMIT_EXCEEDED（账号级限流）：
 	// retryDelay >= 阈值：直接限流模型，不重试
 	// 注意：如果上游未提供 retryDelay，parseAntigravitySmartRetryInfo 已设置为默认 30s
 	if info.RetryDelay >= antigravityRateLimitThreshold {
-		return false, true, info.RetryDelay, info.ModelName
+		return false, true, info.RetryDelay, info.ModelName, false
 	}
 
 	// retryDelay < 阈值：智能重试
@@ -2302,7 +2479,7 @@ func shouldTriggerAntigravitySmartRetry(account *Account, respBody []byte) (shou
 		waitDuration = antigravitySmartRetryMinWait
 	}
 
-	return true, false, waitDuration, info.ModelName
+	return true, false, waitDuration, info.ModelName, false
 }
 
 // handleModelRateLimitParams 模型级限流处理参数
@@ -2328,8 +2505,9 @@ type handleModelRateLimitResult struct {
 
 // handleModelRateLimit 处理模型级限流（在原有逻辑之前调用）
 // 仅处理 429/503，解析模型名和 retryDelay
-// - retryDelay < antigravityRateLimitThreshold: 返回 ShouldRetry=true，由调用方等待后重试
-// - retryDelay >= antigravityRateLimitThreshold: 设置模型限流 + 清除粘性会话 + 返回 SwitchError
+// - MODEL_CAPACITY_EXHAUSTED: 返回 Handled=true（实际重试由 handleSmartRetry 处理）
+// - RATE_LIMIT_EXCEEDED + retryDelay < 阈值: 返回 ShouldRetry=true，由调用方等待后重试
+// - RATE_LIMIT_EXCEEDED + retryDelay >= 阈值: 设置模型限流 + 清除粘性会话 + 返回 SwitchError
 func (s *AntigravityGatewayService) handleModelRateLimit(p *handleModelRateLimitParams) *handleModelRateLimitResult {
 	if p.statusCode != 429 && p.statusCode != 503 {
 		return &handleModelRateLimitResult{Handled: false}
@@ -2340,7 +2518,17 @@ func (s *AntigravityGatewayService) handleModelRateLimit(p *handleModelRateLimit
 		return &handleModelRateLimitResult{Handled: false}
 	}
 
-	// < antigravityRateLimitThreshold: 等待后重试
+	// MODEL_CAPACITY_EXHAUSTED：模型容量不足，所有账号共享同一容量池
+	// 切换账号无意义，不设置模型限流（实际重试由 handleSmartRetry 处理）
+	if info.IsModelCapacityExhausted {
+		log.Printf("%s status=%d model_capacity_exhausted model=%s (not switching account, retry handled by smart retry)",
+			p.prefix, p.statusCode, info.ModelName)
+		return &handleModelRateLimitResult{
+			Handled: true,
+		}
+	}
+
+	// RATE_LIMIT_EXCEEDED: < antigravityRateLimitThreshold: 等待后重试
 	if info.RetryDelay < antigravityRateLimitThreshold {
 		logger.LegacyPrintf("service.antigravity_gateway", "%s status=%d model_rate_limit_wait model=%s wait=%v",
 			p.prefix, p.statusCode, info.ModelName, info.RetryDelay)
@@ -2351,7 +2539,7 @@ func (s *AntigravityGatewayService) handleModelRateLimit(p *handleModelRateLimit
 		}
 	}
 
-	// >= antigravityRateLimitThreshold: 设置限流 + 清除粘性会话 + 切换账号
+	// RATE_LIMIT_EXCEEDED: >= antigravityRateLimitThreshold: 设置限流 + 清除粘性会话 + 切换账号
 	s.setModelRateLimitAndClearSession(p, info)
 
 	return &handleModelRateLimitResult{
@@ -2903,9 +3091,14 @@ returnResponse:
 	// 选择最后一个有效响应
 	finalResponse := pickGeminiCollectResult(last, lastWithParts)
 
-	// 处理空响应情况
+	// 处理空响应情况 — 触发同账号重试 + failover 切换账号
 	if last == nil && lastWithParts == nil {
-		logger.LegacyPrintf("service.antigravity_gateway", "[antigravity-Forward] warning: empty stream response, no valid chunks received")
+		logger.LegacyPrintf("service.antigravity_gateway", "[antigravity-Forward] warning: empty stream response (gemini non-stream), triggering failover")
+		return nil, &UpstreamFailoverError{
+			StatusCode:             http.StatusBadGateway,
+			ResponseBody:           []byte(`{"error":"empty stream response from upstream"}`),
+			RetryableOnSameAccount: true,
+		}
 	}
 
 	// 如果收集到了图片 parts，需要合并到最终响应中
@@ -3123,6 +3316,21 @@ func (s *AntigravityGatewayService) writeMappedClaudeError(c *gin.Context, accou
 		logger.LegacyPrintf("service.antigravity_gateway", "[antigravity-Forward] upstream_error status=%d body=%s", upstreamStatus, truncateForLog(body, maxBytes))
 	}
 
+	// 检查错误透传规则
+	if ptStatus, ptErrType, ptErrMsg, matched := applyErrorPassthroughRule(
+		c, account.Platform, upstreamStatus, body,
+		0, "", "",
+	); matched {
+		c.JSON(ptStatus, gin.H{
+			"type":  "error",
+			"error": gin.H{"type": ptErrType, "message": ptErrMsg},
+		})
+		if upstreamMsg == "" {
+			return fmt.Errorf("upstream error: %d", upstreamStatus)
+		}
+		return fmt.Errorf("upstream error: %d message=%s", upstreamStatus, upstreamMsg)
+	}
+
 	var statusCode int
 	var errType, errMsg string
 
@@ -3320,10 +3528,14 @@ returnResponse:
 	// 选择最后一个有效响应
 	finalResponse := pickGeminiCollectResult(last, lastWithParts)
 
-	// 处理空响应情况
+	// 处理空响应情况 — 触发同账号重试 + failover 切换账号
 	if last == nil && lastWithParts == nil {
-		logger.LegacyPrintf("service.antigravity_gateway", "[antigravity-Forward] warning: empty stream response, no valid chunks received")
-		return nil, s.writeClaudeError(c, http.StatusBadGateway, "upstream_error", "Empty response from upstream")
+		logger.LegacyPrintf("service.antigravity_gateway", "[antigravity-Forward] warning: empty stream response (claude non-stream), triggering failover")
+		return nil, &UpstreamFailoverError{
+			StatusCode:             http.StatusBadGateway,
+			ResponseBody:           []byte(`{"error":"empty stream response from upstream"}`),
+			RetryableOnSameAccount: true,
+		}
 	}
 
 	// 将收集的所有 parts 合并到最终响应中

backend/internal/service/antigravity_gateway_service_test.go

@@ -592,6 +592,75 @@ func TestHandleClaudeStreamingResponse_NormalComplete(t *testing.T) {
 	require.NotContains(t, body, "event: error")
 }
 
+// TestHandleGeminiStreamingResponse_ThoughtsTokenCount
+// 验证：Gemini 流式转发时 thoughtsTokenCount 被计入 OutputTokens
+func TestHandleGeminiStreamingResponse_ThoughtsTokenCount(t *testing.T) {
+	gin.SetMode(gin.TestMode)
+	svc := newAntigravityTestService(&config.Config{
+		Gateway: config.GatewayConfig{MaxLineSize: defaultMaxLineSize},
+	})
+
+	rec := httptest.NewRecorder()
+	c, _ := gin.CreateTestContext(rec)
+	c.Request = httptest.NewRequest(http.MethodPost, "/", nil)
+
+	pr, pw := io.Pipe()
+	resp := &http.Response{StatusCode: http.StatusOK, Body: pr, Header: http.Header{}}
+
+	go func() {
+		defer func() { _ = pw.Close() }()
+		fmt.Fprintln(pw, `data: {"candidates":[{"content":{"parts":[{"text":"Hello"}]}}],"usageMetadata":{"promptTokenCount":100,"candidatesTokenCount":20,"thoughtsTokenCount":50}}`)
+		fmt.Fprintln(pw, "")
+		fmt.Fprintln(pw, `data: {"candidates":[{"content":{"parts":[{"text":" world"}]},"finishReason":"STOP"}],"usageMetadata":{"promptTokenCount":100,"candidatesTokenCount":30,"thoughtsTokenCount":80,"cachedContentTokenCount":10}}`)
+		fmt.Fprintln(pw, "")
+	}()
+
+	result, err := svc.handleGeminiStreamingResponse(c, resp, time.Now())
+	_ = pr.Close()
+
+	require.NoError(t, err)
+	require.NotNil(t, result)
+	require.NotNil(t, result.usage)
+	// promptTokenCount=100, cachedContentTokenCount=10 → InputTokens=90
+	require.Equal(t, 90, result.usage.InputTokens)
+	// candidatesTokenCount=30 + thoughtsTokenCount=80 → OutputTokens=110
+	require.Equal(t, 110, result.usage.OutputTokens)
+	require.Equal(t, 10, result.usage.CacheReadInputTokens)
+}
+
+// TestHandleClaudeStreamingResponse_ThoughtsTokenCount
+// 验证：Gemini→Claude 流式转换时 thoughtsTokenCount 被计入 OutputTokens
+func TestHandleClaudeStreamingResponse_ThoughtsTokenCount(t *testing.T) {
+	gin.SetMode(gin.TestMode)
+	svc := newAntigravityTestService(&config.Config{
+		Gateway: config.GatewayConfig{MaxLineSize: defaultMaxLineSize},
+	})
+
+	rec := httptest.NewRecorder()
+	c, _ := gin.CreateTestContext(rec)
+	c.Request = httptest.NewRequest(http.MethodPost, "/", nil)
+
+	pr, pw := io.Pipe()
+	resp := &http.Response{StatusCode: http.StatusOK, Body: pr, Header: http.Header{}}
+
+	go func() {
+		defer func() { _ = pw.Close() }()
+		fmt.Fprintln(pw, `data: {"response":{"candidates":[{"content":{"parts":[{"text":"Hi"}]},"finishReason":"STOP"}],"usageMetadata":{"promptTokenCount":50,"candidatesTokenCount":10,"thoughtsTokenCount":25}}}`)
+		fmt.Fprintln(pw, "")
+	}()
+
+	result, err := svc.handleClaudeStreamingResponse(c, resp, time.Now(), "gemini-2.5-pro")
+	_ = pr.Close()
+
+	require.NoError(t, err)
+	require.NotNil(t, result)
+	require.NotNil(t, result.usage)
+	// promptTokenCount=50 → InputTokens=50
+	require.Equal(t, 50, result.usage.InputTokens)
+	// candidatesTokenCount=10 + thoughtsTokenCount=25 → OutputTokens=35
+	require.Equal(t, 35, result.usage.OutputTokens)
+}
+
 // --- 流式客户端断开检测测试 ---
 
 // TestStreamUpstreamResponse_ClientDisconnectDrainsUsage

backend/internal/service/antigravity_oauth_service.go

@@ -192,6 +192,43 @@ func (s *AntigravityOAuthService) RefreshToken(ctx context.Context, refreshToken
 	return nil, fmt.Errorf("token 刷新失败 (重试后): %w", lastErr)
 }
 
+// ValidateRefreshToken 用 refresh token 验证并获取完整的 token 信息（含 email 和 project_id）
+func (s *AntigravityOAuthService) ValidateRefreshToken(ctx context.Context, refreshToken string, proxyID *int64) (*AntigravityTokenInfo, error) {
+	var proxyURL string
+	if proxyID != nil {
+		proxy, err := s.proxyRepo.GetByID(ctx, *proxyID)
+		if err == nil && proxy != nil {
+			proxyURL = proxy.URL()
+		}
+	}
+
+	// 刷新 token
+	tokenInfo, err := s.RefreshToken(ctx, refreshToken, proxyURL)
+	if err != nil {
+		return nil, err
+	}
+
+	// 获取用户信息（email）
+	client := antigravity.NewClient(proxyURL)
+	userInfo, err := client.GetUserInfo(ctx, tokenInfo.AccessToken)
+	if err != nil {
+		fmt.Printf("[AntigravityOAuth] 警告: 获取用户信息失败: %v\n", err)
+	} else {
+		tokenInfo.Email = userInfo.Email
+	}
+
+	// 获取 project_id（容错，失败不阻塞）
+	projectID, loadErr := s.loadProjectIDWithRetry(ctx, tokenInfo.AccessToken, proxyURL, 3)
+	if loadErr != nil {
+		fmt.Printf("[AntigravityOAuth] 警告: 获取 project_id 失败（重试后）: %v\n", loadErr)
+		tokenInfo.ProjectIDMissing = true
+	} else {
+		tokenInfo.ProjectID = projectID
+	}
+
+	return tokenInfo, nil
+}
+
 func isNonRetryableAntigravityOAuthError(err error) bool {
 	msg := err.Error()
 	nonRetryable := []string{
@@ -273,12 +310,21 @@ func (s *AntigravityOAuthService) loadProjectIDWithRetry(ctx context.Context, ac
 		}
 
 		client := antigravity.NewClient(proxyURL)
-		loadResp, _, err := client.LoadCodeAssist(ctx, accessToken)
+		loadResp, loadRaw, err := client.LoadCodeAssist(ctx, accessToken)
 
 		if err == nil && loadResp != nil && loadResp.CloudAICompanionProject != "" {
 			return loadResp.CloudAICompanionProject, nil
 		}
 
+		if err == nil {
+			if projectID, onboardErr := tryOnboardProjectID(ctx, client, accessToken, loadRaw); onboardErr == nil && projectID != "" {
+				return projectID, nil
+			} else if onboardErr != nil {
+				lastErr = onboardErr
+				continue
+			}
+		}
+
 		// 记录错误
 		if err != nil {
 			lastErr = err
@@ -292,6 +338,65 @@ func (s *AntigravityOAuthService) loadProjectIDWithRetry(ctx context.Context, ac
 	return "", fmt.Errorf("获取 project_id 失败 (重试 %d 次后): %w", maxRetries, lastErr)
 }
 
+func tryOnboardProjectID(ctx context.Context, client *antigravity.Client, accessToken string, loadRaw map[string]any) (string, error) {
+	tierID := resolveDefaultTierID(loadRaw)
+	if tierID == "" {
+		return "", fmt.Errorf("loadCodeAssist 未返回可用的默认 tier")
+	}
+
+	projectID, err := client.OnboardUser(ctx, accessToken, tierID)
+	if err != nil {
+		return "", fmt.Errorf("onboardUser 失败 (tier=%s): %w", tierID, err)
+	}
+	return projectID, nil
+}
+
+func resolveDefaultTierID(loadRaw map[string]any) string {
+	if len(loadRaw) == 0 {
+		return ""
+	}
+
+	rawTiers, ok := loadRaw["allowedTiers"]
+	if !ok {
+		return ""
+	}
+
+	tiers, ok := rawTiers.([]any)
+	if !ok {
+		return ""
+	}
+
+	for _, rawTier := range tiers {
+		tier, ok := rawTier.(map[string]any)
+		if !ok {
+			continue
+		}
+		if isDefault, _ := tier["isDefault"].(bool); !isDefault {
+			continue
+		}
+		if id, ok := tier["id"].(string); ok {
+			id = strings.TrimSpace(id)
+			if id != "" {
+				return id
+			}
+		}
+	}
+
+	return ""
+}
+
+// FillProjectID 仅获取 project_id，不刷新 OAuth token
+func (s *AntigravityOAuthService) FillProjectID(ctx context.Context, account *Account, accessToken string) (string, error) {
+	var proxyURL string
+	if account.ProxyID != nil {
+		proxy, err := s.proxyRepo.GetByID(ctx, *account.ProxyID)
+		if err == nil && proxy != nil {
+			proxyURL = proxy.URL()
+		}
+	}
+	return s.loadProjectIDWithRetry(ctx, accessToken, proxyURL, 3)
+}
+
 // BuildAccountCredentials 构建账户凭证
 func (s *AntigravityOAuthService) BuildAccountCredentials(tokenInfo *AntigravityTokenInfo) map[string]any {
 	creds := map[string]any{

backend/internal/service/antigravity_oauth_service_test.go

+package service
+
+import (
+	"testing"
+)
+
+func TestResolveDefaultTierID(t *testing.T) {
+	t.Parallel()
+
+	tests := []struct {
+		name    string
+		loadRaw map[string]any
+		want    string
+	}{
+		{
+			name:    "nil loadRaw",
+			loadRaw: nil,
+			want:    "",
+		},
+		{
+			name: "missing allowedTiers",
+			loadRaw: map[string]any{
+				"paidTier": map[string]any{"id": "g1-pro-tier"},
+			},
+			want: "",
+		},
+		{
+			name:    "empty allowedTiers",
+			loadRaw: map[string]any{"allowedTiers": []any{}},
+			want:    "",
+		},
+		{
+			name: "tier missing id field",
+			loadRaw: map[string]any{
+				"allowedTiers": []any{
+					map[string]any{"isDefault": true},
+				},
+			},
+			want: "",
+		},
+		{
+			name: "allowedTiers but no default",
+			loadRaw: map[string]any{
+				"allowedTiers": []any{
+					map[string]any{"id": "free-tier", "isDefault": false},
+					map[string]any{"id": "standard-tier", "isDefault": false},
+				},
+			},
+			want: "",
+		},
+		{
+			name: "default tier found",
+			loadRaw: map[string]any{
+				"allowedTiers": []any{
+					map[string]any{"id": "free-tier", "isDefault": true},
+					map[string]any{"id": "standard-tier", "isDefault": false},
+				},
+			},
+			want: "free-tier",
+		},
+		{
+			name: "default tier id with spaces",
+			loadRaw: map[string]any{
+				"allowedTiers": []any{
+					map[string]any{"id": "  standard-tier  ", "isDefault": true},
+				},
+			},
+			want: "standard-tier",
+		},
+	}
+
+	for _, tc := range tests {
+		t.Run(tc.name, func(t *testing.T) {
+			t.Parallel()
+
+			got := resolveDefaultTierID(tc.loadRaw)
+			if got != tc.want {
+				t.Fatalf("resolveDefaultTierID() = %q, want %q", got, tc.want)
+			}
+		})
+	}
+}

backend/internal/service/antigravity_rate_limit_test.go

@@ -92,7 +92,9 @@ func (s *stubAntigravityAccountRepo) SetModelRateLimit(ctx context.Context, id i
 	return nil
 }
 
-func TestAntigravityRetryLoop_URLFallback_UsesLatestSuccess(t *testing.T) {
+func TestAntigravityRetryLoop_NoURLFallback_UsesConfiguredBaseURL(t *testing.T) {
+	t.Setenv(antigravityForwardBaseURLEnv, "")
+
 	oldBaseURLs := append([]string(nil), antigravity.BaseURLs...)
 	oldAvailability := antigravity.DefaultURLAvailability
 	defer func() {
@@ -137,15 +139,16 @@ func TestAntigravityRetryLoop_URLFallback_UsesLatestSuccess(t *testing.T) {
 	require.NotNil(t, result)
 	require.NotNil(t, result.resp)
 	defer func() { _ = result.resp.Body.Close() }()
-	require.Equal(t, http.StatusOK, result.resp.StatusCode)
-	require.False(t, handleErrorCalled)
-	require.Len(t, upstream.calls, 2)
-	require.True(t, strings.HasPrefix(upstream.calls[0], base1))
-	require.True(t, strings.HasPrefix(upstream.calls[1], base2))
+	require.Equal(t, http.StatusTooManyRequests, result.resp.StatusCode)
+	require.True(t, handleErrorCalled)
+	require.Len(t, upstream.calls, antigravityMaxRetries)
+	for _, callURL := range upstream.calls {
+		require.True(t, strings.HasPrefix(callURL, base1))
+	}
 
 	available := antigravity.DefaultURLAvailability.GetAvailableURLs()
 	require.NotEmpty(t, available)
-	require.Equal(t, base2, available[0])
+	require.Equal(t, base1, available[0])
 }
 
 // TestHandleUpstreamError_429_ModelRateLimit 测试 429 模型限流场景
@@ -194,13 +197,14 @@ func TestHandleUpstreamError_429_NonModelRateLimit(t *testing.T) {
 	require.Equal(t, "claude-sonnet-4-5", repo.modelRateLimitCalls[0].modelKey)
 }
 
-// TestHandleUpstreamError_503_ModelRateLimit 测试 503 模型限流场景
-func TestHandleUpstreamError_503_ModelRateLimit(t *testing.T) {
+// TestHandleUpstreamError_503_ModelCapacityExhausted 测试 503 模型容量不足场景
+// MODEL_CAPACITY_EXHAUSTED 时应等待重试，不切换账号
+func TestHandleUpstreamError_503_ModelCapacityExhausted(t *testing.T) {
 	repo := &stubAntigravityAccountRepo{}
 	svc := &AntigravityGatewayService{accountRepo: repo}
 	account := &Account{ID: 3, Name: "acc-3", Platform: PlatformAntigravity}
 
-	// 503 + MODEL_CAPACITY_EXHAUSTED → 模型限流
+	// 503 + MODEL_CAPACITY_EXHAUSTED → 等待重试，不切换账号
 	body := []byte(`{
 		"error": {
 			"status": "UNAVAILABLE",
@@ -213,13 +217,13 @@ func TestHandleUpstreamError_503_ModelRateLimit(t *testing.T) {
 
 	result := svc.handleUpstreamError(context.Background(), "[test]", account, http.StatusServiceUnavailable, http.Header{}, body, "gemini-3-pro-high", 0, "", false)
 
-	// 应该触发模型限流
+	// MODEL_CAPACITY_EXHAUSTED 应该标记为已处理，不切换账号，不设置模型限流
+	// 实际重试由 handleSmartRetry 处理
 	require.NotNil(t, result)
 	require.True(t, result.Handled)
-	require.NotNil(t, result.SwitchError)
-	require.Equal(t, "gemini-3-pro-high", result.SwitchError.RateLimitedModel)
-	require.Len(t, repo.modelRateLimitCalls, 1)
-	require.Equal(t, "gemini-3-pro-high", repo.modelRateLimitCalls[0].modelKey)
+	require.False(t, result.ShouldRetry, "MODEL_CAPACITY_EXHAUSTED should not trigger retry from handleModelRateLimit path")
+	require.Nil(t, result.SwitchError, "MODEL_CAPACITY_EXHAUSTED should not trigger account switch")
+	require.Empty(t, repo.modelRateLimitCalls, "MODEL_CAPACITY_EXHAUSTED should not set model rate limit")
 }
 
 // TestHandleUpstreamError_503_NonModelRateLimit 测试 503 非模型限流场景（不处理）
@@ -307,11 +311,12 @@ func TestParseGeminiRateLimitResetTime_QuotaResetDelay_RoundsUp(t *testing.T) {
 
 func TestParseAntigravitySmartRetryInfo(t *testing.T) {
 	tests := []struct {
-		name          string
-		body          string
-		expectedDelay time.Duration
-		expectedModel string
-		expectedNil   bool
+		name                             string
+		body                             string
+		expectedDelay                    time.Duration
+		expectedModel                    string
+		expectedNil                      bool
+		expectedIsModelCapacityExhausted bool
 	}{
 		{
 			name: "valid complete response with RATE_LIMIT_EXCEEDED",
@@ -374,8 +379,9 @@ func TestParseAntigravitySmartRetryInfo(t *testing.T) {
 					"message": "No capacity available for model gemini-3-pro-high on the server"
 				}
 			}`,
-			expectedDelay: 39 * time.Second,
-			expectedModel: "gemini-3-pro-high",
+			expectedDelay:                    39 * time.Second,
+			expectedModel:                    "gemini-3-pro-high",
+			expectedIsModelCapacityExhausted: true,
 		},
 		{
 			name: "503 UNAVAILABLE without MODEL_CAPACITY_EXHAUSTED - should return nil",
@@ -486,6 +492,9 @@ func TestParseAntigravitySmartRetryInfo(t *testing.T) {
 			if result.ModelName != tt.expectedModel {
 				t.Errorf("ModelName = %q, want %q", result.ModelName, tt.expectedModel)
 			}
+			if result.IsModelCapacityExhausted != tt.expectedIsModelCapacityExhausted {
+				t.Errorf("IsModelCapacityExhausted = %v, want %v", result.IsModelCapacityExhausted, tt.expectedIsModelCapacityExhausted)
+			}
 		})
 	}
 }
@@ -497,13 +506,14 @@ func TestShouldTriggerAntigravitySmartRetry(t *testing.T) {
 	apiKeyAccount := &Account{Type: AccountTypeAPIKey}
 
 	tests := []struct {
-		name                    string
-		account                 *Account
-		body                    string
-		expectedShouldRetry     bool
-		expectedShouldRateLimit bool
-		minWait                 time.Duration
-		modelName               string
+		name                             string
+		account                          *Account
+		body                             string
+		expectedShouldRetry              bool
+		expectedShouldRateLimit          bool
+		expectedIsModelCapacityExhausted bool
+		minWait                          time.Duration
+		modelName                        string
 	}{
 		{
 			name:    "OAuth account with short delay (< 7s) - smart retry",
@@ -617,13 +627,14 @@ func TestShouldTriggerAntigravitySmartRetry(t *testing.T) {
 					]
 				}
 			}`,
-			expectedShouldRetry:     false,
-			expectedShouldRateLimit: true,
-			minWait:                 39 * time.Second,
-			modelName:               "gemini-3-pro-high",
+			expectedShouldRetry:              true,
+			expectedShouldRateLimit:          false,
+			expectedIsModelCapacityExhausted: true,
+			minWait:                          1 * time.Second,
+			modelName:                        "gemini-3-pro-high",
 		},
 		{
-			name:    "503 UNAVAILABLE with MODEL_CAPACITY_EXHAUSTED - no retryDelay - use default rate limit",
+			name:    "503 UNAVAILABLE with MODEL_CAPACITY_EXHAUSTED - no retryDelay - use fixed wait",
 			account: oauthAccount,
 			body: `{
 				"error": {
@@ -635,10 +646,11 @@ func TestShouldTriggerAntigravitySmartRetry(t *testing.T) {
 					"message": "No capacity available for model gemini-2.5-flash on the server"
 				}
 			}`,
-			expectedShouldRetry:     false,
-			expectedShouldRateLimit: true,
-			minWait:                 30 * time.Second,
-			modelName:               "gemini-2.5-flash",
+			expectedShouldRetry:              true,
+			expectedShouldRateLimit:          false,
+			expectedIsModelCapacityExhausted: true,
+			minWait:                          1 * time.Second,
+			modelName:                        "gemini-2.5-flash",
 		},
 		{
 			name:    "429 RESOURCE_EXHAUSTED with RATE_LIMIT_EXCEEDED - no retryDelay - use default rate limit",
@@ -662,13 +674,16 @@ func TestShouldTriggerAntigravitySmartRetry(t *testing.T) {
 
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {
-			shouldRetry, shouldRateLimit, wait, model := shouldTriggerAntigravitySmartRetry(tt.account, []byte(tt.body))
+			shouldRetry, shouldRateLimit, wait, model, isModelCapacityExhausted := shouldTriggerAntigravitySmartRetry(tt.account, []byte(tt.body))
 			if shouldRetry != tt.expectedShouldRetry {
 				t.Errorf("shouldRetry = %v, want %v", shouldRetry, tt.expectedShouldRetry)
 			}
 			if shouldRateLimit != tt.expectedShouldRateLimit {
 				t.Errorf("shouldRateLimit = %v, want %v", shouldRateLimit, tt.expectedShouldRateLimit)
 			}
+			if isModelCapacityExhausted != tt.expectedIsModelCapacityExhausted {
+				t.Errorf("isModelCapacityExhausted = %v, want %v", isModelCapacityExhausted, tt.expectedIsModelCapacityExhausted)
+			}
 			if shouldRetry {
 				if wait < tt.minWait {
 					t.Errorf("wait = %v, want >= %v", wait, tt.minWait)
@@ -921,6 +936,22 @@ func TestIsAntigravityAccountSwitchError(t *testing.T) {
 	}
 }
 
+func TestResolveAntigravityForwardBaseURL_DefaultDaily(t *testing.T) {
+	t.Setenv(antigravityForwardBaseURLEnv, "")
+
+	oldBaseURLs := append([]string(nil), antigravity.BaseURLs...)
+	defer func() {
+		antigravity.BaseURLs = oldBaseURLs
+	}()
+
+	prodURL := "https://prod.test"
+	dailyURL := "https://daily.test"
+	antigravity.BaseURLs = []string{dailyURL, prodURL}
+
+	resolved := resolveAntigravityForwardBaseURL()
+	require.Equal(t, dailyURL, resolved)
+}
+
 func TestAntigravityAccountSwitchError_Error(t *testing.T) {
 	err := &AntigravityAccountSwitchError{
 		OriginalAccountID: 789,

backend/internal/service/antigravity_single_account_retry_test.go

@@ -153,13 +153,14 @@ func TestHandleSmartRetry_503_LongDelay_NoSingleAccountRetry_StillSwitches(t *te
 		Platform: PlatformAntigravity,
 	}
 
-	// 503 + 39s >= 7s 阈值
+	// 503 + 39s >= 7s 阈值（使用 RATE_LIMIT_EXCEEDED 而非 MODEL_CAPACITY_EXHAUSTED，
+	// 因为 MODEL_CAPACITY_EXHAUSTED 走独立的重试路径，不触发 shouldRateLimitModel）
 	respBody := []byte(`{
 		"error": {
 			"code": 503,
-			"status": "UNAVAILABLE",
+			"status": "RESOURCE_EXHAUSTED",
 			"details": [
-				{"@type": "type.googleapis.com/google.rpc.ErrorInfo", "metadata": {"model": "gemini-3-pro-high"}, "reason": "MODEL_CAPACITY_EXHAUSTED"},
+				{"@type": "type.googleapis.com/google.rpc.ErrorInfo", "metadata": {"model": "gemini-3-pro-high"}, "reason": "RATE_LIMIT_EXCEEDED"},
 				{"@type": "type.googleapis.com/google.rpc.RetryInfo", "retryDelay": "39s"}
 			]
 		}
@@ -339,13 +340,14 @@ func TestHandleSmartRetry_503_ShortDelay_SingleAccountRetry_NoRateLimit(t *testi
 
 // TestHandleSmartRetry_503_ShortDelay_NoSingleAccountRetry_SetsRateLimit
 // 对照组：503 + retryDelay < 7s + 无 SingleAccountRetry → 智能重试耗尽后照常设限流
+// 使用 RATE_LIMIT_EXCEEDED 而非 MODEL_CAPACITY_EXHAUSTED，因为后者走独立的 60 次重试路径
 func TestHandleSmartRetry_503_ShortDelay_NoSingleAccountRetry_SetsRateLimit(t *testing.T) {
 	failRespBody := `{
 		"error": {
 			"code": 503,
-			"status": "UNAVAILABLE",
+			"status": "RESOURCE_EXHAUSTED",
 			"details": [
-				{"@type": "type.googleapis.com/google.rpc.ErrorInfo", "metadata": {"model": "gemini-3-flash"}, "reason": "MODEL_CAPACITY_EXHAUSTED"},
+				{"@type": "type.googleapis.com/google.rpc.ErrorInfo", "metadata": {"model": "gemini-3-flash"}, "reason": "RATE_LIMIT_EXCEEDED"},
 				{"@type": "type.googleapis.com/google.rpc.RetryInfo", "retryDelay": "0.1s"}
 			]
 		}
@@ -371,9 +373,9 @@ func TestHandleSmartRetry_503_ShortDelay_NoSingleAccountRetry_SetsRateLimit(t *t
 	respBody := []byte(`{
 		"error": {
 			"code": 503,
-			"status": "UNAVAILABLE",
+			"status": "RESOURCE_EXHAUSTED",
 			"details": [
-				{"@type": "type.googleapis.com/google.rpc.ErrorInfo", "metadata": {"model": "gemini-3-flash"}, "reason": "MODEL_CAPACITY_EXHAUSTED"},
+				{"@type": "type.googleapis.com/google.rpc.ErrorInfo", "metadata": {"model": "gemini-3-flash"}, "reason": "RATE_LIMIT_EXCEEDED"},
 				{"@type": "type.googleapis.com/google.rpc.RetryInfo", "retryDelay": "0.1s"}
 			]
 		}

backend/internal/service/antigravity_smart_retry_test.go

@@ -294,8 +294,9 @@ func TestHandleSmartRetry_ShortDelay_SmartRetryFailed_ReturnsSwitchError(t *test
 	require.Len(t, upstream.calls, 1, "should have made one retry call (max attempts)")
 }
 
-// TestHandleSmartRetry_503_ModelCapacityExhausted_ReturnsSwitchError 测试 503 MODEL_CAPACITY_EXHAUSTED 返回 switchError
-func TestHandleSmartRetry_503_ModelCapacityExhausted_ReturnsSwitchError(t *testing.T) {
+// TestHandleSmartRetry_503_ModelCapacityExhausted_RetrySuccess 测试 503 MODEL_CAPACITY_EXHAUSTED 重试成功
+// MODEL_CAPACITY_EXHAUSTED 使用固定 1s 间隔重试，不切换账号
+func TestHandleSmartRetry_503_ModelCapacityExhausted_RetrySuccess(t *testing.T) {
 	repo := &stubAntigravityAccountRepo{}
 	account := &Account{
 		ID:       3,
@@ -304,7 +305,7 @@ func TestHandleSmartRetry_503_ModelCapacityExhausted_ReturnsSwitchError(t *testi
 		Platform: PlatformAntigravity,
 	}
 
-	// 503 + MODEL_CAPACITY_EXHAUSTED + 39s >= 7s 阈值
+	// 503 + MODEL_CAPACITY_EXHAUSTED + 39s（上游 retryDelay 应被忽略，使用固定 1s）
 	respBody := []byte(`{
 		"error": {
 			"code": 503,
@@ -322,6 +323,14 @@ func TestHandleSmartRetry_503_ModelCapacityExhausted_ReturnsSwitchError(t *testi
 		Body:       io.NopCloser(bytes.NewReader(respBody)),
 	}
 
+	// mock: 第 1 次重试返回 200 成功
+	upstream := &mockSmartRetryUpstream{
+		responses: []*http.Response{
+			{StatusCode: http.StatusOK, Header: http.Header{}, Body: io.NopCloser(strings.NewReader(`{"ok":true}`))},
+		},
+		errors: []error{nil},
+	}
+
 	params := antigravityRetryLoopParams{
 		ctx:             context.Background(),
 		prefix:          "[test]",
@@ -330,6 +339,7 @@ func TestHandleSmartRetry_503_ModelCapacityExhausted_ReturnsSwitchError(t *testi
 		action:          "generateContent",
 		body:            []byte(`{"input":"test"}`),
 		accountRepo:     repo,
+		httpUpstream:    upstream,
 		isStickySession: true,
 		handleError: func(ctx context.Context, prefix string, account *Account, statusCode int, headers http.Header, body []byte, requestedModel string, groupID int64, sessionHash string, isStickySession bool) *handleModelRateLimitResult {
 			return nil
@@ -343,16 +353,67 @@ func TestHandleSmartRetry_503_ModelCapacityExhausted_ReturnsSwitchError(t *testi
 
 	require.NotNil(t, result)
 	require.Equal(t, smartRetryActionBreakWithResp, result.action)
-	require.Nil(t, result.resp)
+	require.NotNil(t, result.resp, "should return successful response")
+	require.Equal(t, http.StatusOK, result.resp.StatusCode)
 	require.Nil(t, result.err)
-	require.NotNil(t, result.switchError, "should return switchError for 503 model capacity exhausted")
-	require.Equal(t, account.ID, result.switchError.OriginalAccountID)
-	require.Equal(t, "gemini-3-pro-high", result.switchError.RateLimitedModel)
-	require.True(t, result.switchError.IsStickySession)
+	require.Nil(t, result.switchError, "MODEL_CAPACITY_EXHAUSTED should not return switchError")
 
-	// 验证模型限流已设置
-	require.Len(t, repo.modelRateLimitCalls, 1)
-	require.Equal(t, "gemini-3-pro-high", repo.modelRateLimitCalls[0].modelKey)
+	// 不应设置模型限流
+	require.Empty(t, repo.modelRateLimitCalls, "MODEL_CAPACITY_EXHAUSTED should not set model rate limit")
+	require.Len(t, upstream.calls, 1, "should have made one retry call before success")
+}
+
+// TestHandleSmartRetry_503_ModelCapacityExhausted_ContextCancel 测试 MODEL_CAPACITY_EXHAUSTED 上下文取消
+func TestHandleSmartRetry_503_ModelCapacityExhausted_ContextCancel(t *testing.T) {
+	repo := &stubAntigravityAccountRepo{}
+	account := &Account{
+		ID:       3,
+		Name:     "acc-3",
+		Type:     AccountTypeOAuth,
+		Platform: PlatformAntigravity,
+	}
+
+	respBody := []byte(`{
+		"error": {
+			"code": 503,
+			"status": "UNAVAILABLE",
+			"details": [
+				{"@type": "type.googleapis.com/google.rpc.ErrorInfo", "metadata": {"model": "gemini-3-pro-high"}, "reason": "MODEL_CAPACITY_EXHAUSTED"},
+				{"@type": "type.googleapis.com/google.rpc.RetryInfo", "retryDelay": "39s"}
+			]
+		}
+	}`)
+	resp := &http.Response{
+		StatusCode: http.StatusServiceUnavailable,
+		Header:     http.Header{},
+		Body:       io.NopCloser(bytes.NewReader(respBody)),
+	}
+
+	// 立即取消上下文，验证重试循环能正确退出
+	ctx, cancel := context.WithCancel(context.Background())
+	cancel()
+
+	params := antigravityRetryLoopParams{
+		ctx:         ctx,
+		prefix:      "[test]",
+		account:     account,
+		accessToken: "token",
+		action:      "generateContent",
+		body:        []byte(`{"input":"test"}`),
+		accountRepo: repo,
+		handleError: func(ctx context.Context, prefix string, account *Account, statusCode int, headers http.Header, body []byte, requestedModel string, groupID int64, sessionHash string, isStickySession bool) *handleModelRateLimitResult {
+			return nil
+		},
+	}
+
+	svc := &AntigravityGatewayService{}
+	result := svc.handleSmartRetry(params, resp, respBody, "https://ag-1.test", 0, []string{"https://ag-1.test"})
+
+	require.NotNil(t, result)
+	require.Equal(t, smartRetryActionBreakWithResp, result.action)
+	require.Error(t, result.err, "should return context error")
+	require.Nil(t, result.switchError, "should not return switchError on context cancel")
+	require.Empty(t, repo.modelRateLimitCalls, "should not set model rate limit on context cancel")
 }
 
 // TestHandleSmartRetry_NonAntigravityAccount_ContinuesDefaultLogic 测试非 Antigravity 平台账号走默认逻辑
@@ -1129,20 +1190,20 @@ func TestHandleSmartRetry_ShortDelay_NetworkError_StickySession_ClearsSession(t
 }
 
 // TestHandleSmartRetry_ShortDelay_503_StickySession_FailedRetry_ClearsSession
-// 503 + 短延迟 + 粘性会话 + 重试失败 → 清除粘性绑定
+// 429 + 短延迟 + 粘性会话 + 重试失败 → 清除粘性绑定
 func TestHandleSmartRetry_ShortDelay_503_StickySession_FailedRetry_ClearsSession(t *testing.T) {
 	failRespBody := `{
 		"error": {
-			"code": 503,
-			"status": "UNAVAILABLE",
+			"code": 429,
+			"status": "RESOURCE_EXHAUSTED",
 			"details": [
-				{"@type": "type.googleapis.com/google.rpc.ErrorInfo", "metadata": {"model": "gemini-3-pro"}, "reason": "MODEL_CAPACITY_EXHAUSTED"},
+				{"@type": "type.googleapis.com/google.rpc.ErrorInfo", "metadata": {"model": "gemini-3-pro"}, "reason": "RATE_LIMIT_EXCEEDED"},
 				{"@type": "type.googleapis.com/google.rpc.RetryInfo", "retryDelay": "0.5s"}
 			]
 		}
 	}`
 	failResp := &http.Response{
-		StatusCode: http.StatusServiceUnavailable,
+		StatusCode: http.StatusTooManyRequests,
 		Header:     http.Header{},
 		Body:       io.NopCloser(strings.NewReader(failRespBody)),
 	}
@@ -1162,16 +1223,16 @@ func TestHandleSmartRetry_ShortDelay_503_StickySession_FailedRetry_ClearsSession
 
 	respBody := []byte(`{
 		"error": {
-			"code": 503,
-			"status": "UNAVAILABLE",
+			"code": 429,
+			"status": "RESOURCE_EXHAUSTED",
 			"details": [
-				{"@type": "type.googleapis.com/google.rpc.ErrorInfo", "metadata": {"model": "gemini-3-pro"}, "reason": "MODEL_CAPACITY_EXHAUSTED"},
+				{"@type": "type.googleapis.com/google.rpc.ErrorInfo", "metadata": {"model": "gemini-3-pro"}, "reason": "RATE_LIMIT_EXCEEDED"},
 				{"@type": "type.googleapis.com/google.rpc.RetryInfo", "retryDelay": "0.5s"}
 			]
 		}
 	}`)
 	resp := &http.Response{
-		StatusCode: http.StatusServiceUnavailable,
+		StatusCode: http.StatusTooManyRequests,
 		Header:     http.Header{},
 		Body:       io.NopCloser(bytes.NewReader(respBody)),
 	}

backend/internal/service/antigravity_token_provider.go

@@ -7,12 +7,14 @@ import (
 	"log/slog"
 	"strconv"
 	"strings"
+	"sync"
 	"time"
 )
 
 const (
 	antigravityTokenRefreshSkew = 3 * time.Minute
 	antigravityTokenCacheSkew   = 5 * time.Minute
+	antigravityBackfillCooldown = 5 * time.Minute
 )
 
 // AntigravityTokenCache Token 缓存接口（复用 GeminiTokenCache 接口定义）
@@ -23,6 +25,7 @@ type AntigravityTokenProvider struct {
 	accountRepo             AccountRepository
 	tokenCache              AntigravityTokenCache
 	antigravityOAuthService *AntigravityOAuthService
+	backfillCooldown        sync.Map // key: int64 (account.ID) → value: time.Time
 }
 
 func NewAntigravityTokenProvider(
@@ -93,13 +96,7 @@ func (p *AntigravityTokenProvider) GetAccessToken(ctx context.Context, account *
 				if err != nil {
 					return "", err
 				}
-				newCredentials := p.antigravityOAuthService.BuildAccountCredentials(tokenInfo)
-				for k, v := range account.Credentials {
-					if _, exists := newCredentials[k]; !exists {
-						newCredentials[k] = v
-					}
-				}
-				account.Credentials = newCredentials
+				p.mergeCredentials(account, tokenInfo)
 				if updateErr := p.accountRepo.Update(ctx, account); updateErr != nil {
 					log.Printf("[AntigravityTokenProvider] Failed to update account credentials: %v", updateErr)
 				}
@@ -113,6 +110,21 @@ func (p *AntigravityTokenProvider) GetAccessToken(ctx context.Context, account *
 		return "", errors.New("access_token not found in credentials")
 	}
 
+	// 如果账号还没有 project_id，尝试在线补齐，避免请求 daily/sandbox 时出现
+	// "Invalid project resource name projects/"。
+	// 仅调用 loadProjectIDWithRetry，不刷新 OAuth token；带冷却机制防止频繁重试。
+	if strings.TrimSpace(account.GetCredential("project_id")) == "" && p.antigravityOAuthService != nil {
+		if p.shouldAttemptBackfill(account.ID) {
+			p.markBackfillAttempted(account.ID)
+			if projectID, err := p.antigravityOAuthService.FillProjectID(ctx, account, accessToken); err == nil && projectID != "" {
+				account.Credentials["project_id"] = projectID
+				if updateErr := p.accountRepo.Update(ctx, account); updateErr != nil {
+					log.Printf("[AntigravityTokenProvider] project_id 补齐持久化失败: %v", updateErr)
+				}
+			}
+		}
+	}
+
 	// 3. 存入缓存（验证版本后再写入，避免异步刷新任务与请求线程的竞态条件）
 	if p.tokenCache != nil {
 		latestAccount, isStale := CheckTokenVersion(ctx, account, p.accountRepo)
@@ -144,6 +156,31 @@ func (p *AntigravityTokenProvider) GetAccessToken(ctx context.Context, account *
 	return accessToken, nil
 }
 
+// mergeCredentials 将 tokenInfo 构建的凭证合并到 account 中，保留原有未覆盖的字段
+func (p *AntigravityTokenProvider) mergeCredentials(account *Account, tokenInfo *AntigravityTokenInfo) {
+	newCredentials := p.antigravityOAuthService.BuildAccountCredentials(tokenInfo)
+	for k, v := range account.Credentials {
+		if _, exists := newCredentials[k]; !exists {
+			newCredentials[k] = v
+		}
+	}
+	account.Credentials = newCredentials
+}
+
+// shouldAttemptBackfill 检查是否应该尝试补齐 project_id（冷却期内不重复尝试）
+func (p *AntigravityTokenProvider) shouldAttemptBackfill(accountID int64) bool {
+	if v, ok := p.backfillCooldown.Load(accountID); ok {
+		if lastAttempt, ok := v.(time.Time); ok {
+			return time.Since(lastAttempt) > antigravityBackfillCooldown
+		}
+	}
+	return true
+}
+
+func (p *AntigravityTokenProvider) markBackfillAttempted(accountID int64) {
+	p.backfillCooldown.Store(accountID, time.Now())
+}
+
 func AntigravityTokenCacheKey(account *Account) string {
 	projectID := strings.TrimSpace(account.GetCredential("project_id"))
 	if projectID != "" {

backend/internal/service/error_passthrough_runtime.go

@@ -61,6 +61,11 @@ func applyErrorPassthroughRule(
 		errMsg = *rule.CustomMessage
 	}
 
+	// 命中 skip_monitoring 时在 context 中标记，供 ops_error_logger 跳过记录。
+	if rule.SkipMonitoring {
+		c.Set(OpsSkipPassthroughKey, true)
+	}
+
 	// 与现有 failover 场景保持一致：命中规则时统一返回 upstream_error。
 	errType = "upstream_error"
 	return status, errType, errMsg, true