Record flux2klein overnight smoke review

benchmarks/flux2klein/README.md

@@ -36,3 +36,7 @@ Each case folder should keep:
 ## Review rule
 
 Do not call a workflow fully optimal from one benchmark image. Use these batches to decide whether a path is promising enough for follow-up.
+
+## Latest review
+
+- 2026-08-21 overnight single-image smoke: clean full-body studio restage on a pure white background, with good identity retention and a tidy silhouette. Treat it as the current benchmark winner for now, but keep one more smoke test queued before any stronger promotion.

benchmarks/flux2klein/overnight-state.json

@@ -21,9 +21,10 @@
       }
     }
   ],
-  "reviewPending": true,
   "optimizedPaths": [
     "optimized/image-edit/flux2klein-9b-fp8-edit-master"
   ],
-  "verdict": null
+  "reviewedAt": "2026-08-21T05:00:00Z",
+  "reviewPending": false,
+  "verdict": "promising; clean identity retention and white background; keep as benchmark winner and queue one more smoke test"
 }

optimized/notes/flux2klein.md

@@ -48,6 +48,7 @@
   - low-step settings when they are clearly validated
   - better reference/outpaint control if portable
 - Treat the `geceff` multi-reference pack as a donor for future reference-heavy variants, not as the first default promotion.
+- The 2026-08-21 overnight `image-edit-9b-distilled` smoke was clean enough to keep the current `flux2klein-9b-fp8-edit-master` lane as the benchmark winner for now, but it still needs another night of proof before any broader claim of stability.
 
 ## What needs Monday proof