diff --git a/internal/runtime/executor/antigravity_executor.go b/internal/runtime/executor/antigravity_executor.go
index 9ade4fbb..38dc0b84 100644
--- a/internal/runtime/executor/antigravity_executor.go
+++ b/internal/runtime/executor/antigravity_executor.go
@@ -155,7 +155,13 @@ func (e *AntigravityExecutor) Execute(ctx context.Context, auth *cliproxyauth.Au
 				log.Debugf("antigravity executor: rate limited on base url %s, retrying with fallback base url: %s", baseURL, baseURLs[idx+1])
 				continue
 			}
-			err = statusErr{code: httpResp.StatusCode, msg: string(bodyBytes)}
+			sErr := statusErr{code: httpResp.StatusCode, msg: string(bodyBytes)}
+			if httpResp.StatusCode == http.StatusTooManyRequests {
+				if retryAfter, parseErr := parseRetryDelay(bodyBytes); parseErr == nil && retryAfter != nil {
+					sErr.retryAfter = retryAfter
+				}
+			}
+			err = sErr
 			return resp, err
 		}
 
@@ -169,7 +175,13 @@ func (e *AntigravityExecutor) Execute(ctx context.Context, auth *cliproxyauth.Au
 
 	switch {
 	case lastStatus != 0:
-		err = statusErr{code: lastStatus, msg: string(lastBody)}
+		sErr := statusErr{code: lastStatus, msg: string(lastBody)}
+		if lastStatus == http.StatusTooManyRequests {
+			if retryAfter, parseErr := parseRetryDelay(lastBody); parseErr == nil && retryAfter != nil {
+				sErr.retryAfter = retryAfter
+			}
+		}
+		err = sErr
 	case lastErr != nil:
 		err = lastErr
 	default:
@@ -259,7 +271,13 @@ func (e *AntigravityExecutor) executeClaudeNonStream(ctx context.Context, auth *
 				log.Debugf("antigravity executor: rate limited on base url %s, retrying with fallback base url: %s", baseURL, baseURLs[idx+1])
 				continue
 			}
-			err = statusErr{code: httpResp.StatusCode, msg: string(bodyBytes)}
+			sErr := statusErr{code: httpResp.StatusCode, msg: string(bodyBytes)}
+			if httpResp.StatusCode == http.StatusTooManyRequests {
+				if retryAfter, parseErr := parseRetryDelay(bodyBytes); parseErr == nil && retryAfter != nil {
+					sErr.retryAfter = retryAfter
+				}
+			}
+			err = sErr
 			return resp, err
 		}
 
@@ -324,7 +342,13 @@ func (e *AntigravityExecutor) executeClaudeNonStream(ctx context.Context, auth *
 
 	switch {
 	case lastStatus != 0:
-		err = statusErr{code: lastStatus, msg: string(lastBody)}
+		sErr := statusErr{code: lastStatus, msg: string(lastBody)}
+		if lastStatus == http.StatusTooManyRequests {
+			if retryAfter, parseErr := parseRetryDelay(lastBody); parseErr == nil && retryAfter != nil {
+				sErr.retryAfter = retryAfter
+			}
+		}
+		err = sErr
 	case lastErr != nil:
 		err = lastErr
 	default:
@@ -599,7 +623,13 @@ func (e *AntigravityExecutor) ExecuteStream(ctx context.Context, auth *cliproxya
 				log.Debugf("antigravity executor: rate limited on base url %s, retrying with fallback base url: %s", baseURL, baseURLs[idx+1])
 				continue
 			}
-			err = statusErr{code: httpResp.StatusCode, msg: string(bodyBytes)}
+			sErr := statusErr{code: httpResp.StatusCode, msg: string(bodyBytes)}
+			if httpResp.StatusCode == http.StatusTooManyRequests {
+				if retryAfter, parseErr := parseRetryDelay(bodyBytes); parseErr == nil && retryAfter != nil {
+					sErr.retryAfter = retryAfter
+				}
+			}
+			err = sErr
 			return nil, err
 		}
 
@@ -654,7 +684,13 @@ func (e *AntigravityExecutor) ExecuteStream(ctx context.Context, auth *cliproxya
 
 	switch {
 	case lastStatus != 0:
-		err = statusErr{code: lastStatus, msg: string(lastBody)}
+		sErr := statusErr{code: lastStatus, msg: string(lastBody)}
+		if lastStatus == http.StatusTooManyRequests {
+			if retryAfter, parseErr := parseRetryDelay(lastBody); parseErr == nil && retryAfter != nil {
+				sErr.retryAfter = retryAfter
+			}
+		}
+		err = sErr
 	case lastErr != nil:
 		err = lastErr
 	default:
@@ -795,12 +831,24 @@ func (e *AntigravityExecutor) CountTokens(ctx context.Context, auth *cliproxyaut
 			log.Debugf("antigravity executor: rate limited on base url %s, retrying with fallback base url: %s", baseURL, baseURLs[idx+1])
 			continue
 		}
-		return cliproxyexecutor.Response{}, statusErr{code: httpResp.StatusCode, msg: string(bodyBytes)}
+		sErr := statusErr{code: httpResp.StatusCode, msg: string(bodyBytes)}
+		if httpResp.StatusCode == http.StatusTooManyRequests {
+			if retryAfter, parseErr := parseRetryDelay(bodyBytes); parseErr == nil && retryAfter != nil {
+				sErr.retryAfter = retryAfter
+			}
+		}
+		return cliproxyexecutor.Response{}, sErr
 	}
 
 	switch {
 	case lastStatus != 0:
-		return cliproxyexecutor.Response{}, statusErr{code: lastStatus, msg: string(lastBody)}
+		sErr := statusErr{code: lastStatus, msg: string(lastBody)}
+		if lastStatus == http.StatusTooManyRequests {
+			if retryAfter, parseErr := parseRetryDelay(lastBody); parseErr == nil && retryAfter != nil {
+				sErr.retryAfter = retryAfter
+			}
+		}
+		return cliproxyexecutor.Response{}, sErr
 	case lastErr != nil:
 		return cliproxyexecutor.Response{}, lastErr
 	default:
@@ -963,7 +1011,13 @@ func (e *AntigravityExecutor) refreshToken(ctx context.Context, auth *cliproxyau
 	}
 
 	if httpResp.StatusCode < http.StatusOK || httpResp.StatusCode >= http.StatusMultipleChoices {
-		return auth, statusErr{code: httpResp.StatusCode, msg: string(bodyBytes)}
+		sErr := statusErr{code: httpResp.StatusCode, msg: string(bodyBytes)}
+		if httpResp.StatusCode == http.StatusTooManyRequests {
+			if retryAfter, parseErr := parseRetryDelay(bodyBytes); parseErr == nil && retryAfter != nil {
+				sErr.retryAfter = retryAfter
+			}
+		}
+		return auth, sErr
 	}
 
 	var tokenResp struct {