diff --git a/litellm/router.py b/litellm/router.py
index e42bc2af398..85586ccab17 100644
--- a/litellm/router.py
+++ b/litellm/router.py
@@ -7572,12 +7572,12 @@ class Router:
             ## LOGGING
             if num_retries > 0:
                 kwargs = self.log_retry(kwargs=kwargs, e=original_exception)
-                skipped_deployment_ids: Final = self._deployment_ids_to_skip_on_retry(
+                first_skipped_ids: Final = self._deployment_ids_to_skip_on_retry(
                     exception=original_exception,
                     already_skipped=kwargs.get("_retry_skipped_deployment_ids"),
                 )
-                if skipped_deployment_ids:
-                    kwargs["_retry_skipped_deployment_ids"] = skipped_deployment_ids
+                if first_skipped_ids:
+                    kwargs["_retry_skipped_deployment_ids"] = first_skipped_ids  # rebind-ok: the next attempt reads it
             else:
                 raise
 
@@ -7647,12 +7647,12 @@ class Router:
                         except Exception:
                             raise e
 
-                    retry_skipped_deployment_ids = self._deployment_ids_to_skip_on_retry(
+                    skipped_ids = self._deployment_ids_to_skip_on_retry(
                         exception=e,
                         already_skipped=kwargs.get("_retry_skipped_deployment_ids"),
                     )
-                    if retry_skipped_deployment_ids:
-                        kwargs["_retry_skipped_deployment_ids"] = retry_skipped_deployment_ids
+                    if skipped_ids:
+                        kwargs["_retry_skipped_deployment_ids"] = skipped_ids  # rebind-ok: the next attempt reads it
                     _timeout = self._time_to_sleep_before_retry(
                         e=e,
                         remaining_retries=remaining_retries,
diff --git a/litellm/utils.py b/litellm/utils.py
index 238225eff99..f99d7e6a4b7 100644
--- a/litellm/utils.py
+++ b/litellm/utils.py
@@ -4919,10 +4919,12 @@ def get_excluded_filtered_deployments(
     across the remaining deployments in the same model group after one of them
     has failed.
 
-    If the filter would leave no deployments, an empty list is returned so the
-    caller raises its usual no-deployments error and the weighted-failover
-    helper falls through to the cross-group fallback path. Returning the
-    original unfiltered list here would re-include the just-failed deployment.
+    If the filter would leave no deployments, an empty list is returned and the
+    caller decides what that means. Weighted failover lets it raise the usual
+    no-deployments error and fall through to the cross-group fallback path; the
+    retry skip in `async_get_healthy_deployments` deliberately falls back to the
+    unfiltered list, so a request every deployment refused still comes back with
+    the provider's own error rather than a no-deployments one.
     """
     if not excluded_deployment_ids:
         return healthy_deployments
diff --git a/tests/e2e/coverage_registry/reliability.yaml b/tests/e2e/coverage_registry/reliability.yaml
index b50551ec105..6b69677d490 100644
--- a/tests/e2e/coverage_registry/reliability.yaml
+++ b/tests/e2e/coverage_registry/reliability.yaml
@@ -7,7 +7,7 @@
 - {id: reliability.retry.timeout.succeeds_within_retries, module: reliability, tier: P0, behavior: retry, variant: timeout, assertions: [succeeds_within_retries], exercised_on: [chat_completions, messages], source: "get_retry_from_policy.py:44", rationale: "Timeout retried per policy"}
 - {id: reliability.retry.429.succeeds_within_retries, module: reliability, tier: P0, behavior: retry, variant: "429", assertions: [succeeds_within_retries], exercised_on: [chat_completions, messages], source: "get_retry_from_policy.py:46", rationale: "429 retried per RateLimitErrorRetries policy"}
 - {id: reliability.retry.auth.succeeds_within_retries, module: reliability, tier: P1, behavior: retry, variant: auth, assertions: [succeeds_within_retries], exercised_on: [chat_completions, messages], source: "get_retry_from_policy.py:42", rationale: "Transient auth glitch retry"}
-- {id: reliability.retry.context_window.succeeds_within_retries, module: reliability, tier: P1, behavior: retry, variant: context_window, assertions: [succeeds_within_retries], exercised_on: [chat_completions, messages], source: "get_retry_from_policy.py:51", rationale: "Multi-attempt on context error"}
+- {id: reliability.retry.context_window.succeeds_within_retries, module: reliability, tier: P1, behavior: retry, variant: context_window, assertions: [succeeds_within_retries], exercised_on: [chat_completions, messages], source: "get_retry_from_policy.py:51", fail_before_fix: proven, rationale: "A context-window 400 under BadRequestErrorRetries retries onto a sibling deployment in the same model group, instead of coming straight back as the 400 the deployment that just refused it returned"}
 - {id: reliability.cooldown.5xx.trips_then_recovers, module: reliability, tier: P0, behavior: cooldown, variant: "5xx", assertions: [trips_then_recovers], exercised_on: [chat_completions, messages], source: "cooldown_handlers.py:40", rationale: "Deployment cools after repeated 5xx, recovers after cooldown_time"}
 - {id: reliability.cooldown.429.trips_then_recovers, module: reliability, tier: P0, behavior: cooldown, variant: "429", assertions: [trips_then_recovers], exercised_on: [chat_completions, messages], source: "cooldown_handlers.py:69", rationale: "Cools on 429, avoids hammering exhausted provider"}
 - {id: reliability.cooldown.auth.trips_then_recovers, module: reliability, tier: P1, behavior: cooldown, variant: auth, assertions: [trips_then_recovers], exercised_on: [chat_completions, messages], source: "cooldown_handlers.py:74", rationale: "Cools on 401 auth error"}