Add transparent cross-family comparison

README.md

@@ -35,6 +35,23 @@ a clean validation split. They are engineering results, not a multi-seed
 scientific claim. The held-out test split was not read during training or
 model selection.
 
+## YOLOv8 and YOLOv26 comparison
+
+Both model families use the same 640 px WALDO 4.0 class contract and clean
+validation split. The YOLOv26 lane uses Obnubilated training only. Its values
+remain pending until each full training run and terminal evaluation complete;
+no cross-family delta is reported early.
+
+| Size | YOLOv8 Standard mAP50-95 | YOLOv8 Obnubilated mAP50-95 | YOLOv26 Obnubilated mAP50-95 | Cross-family status |
+| --- | ---: | ---: | ---: | --- |
+| N | 0.4425 | 0.4377 | Pending | Awaiting YOLOv26 terminal evaluation |
+| S | 0.4978 | 0.5002 | Pending | Awaiting YOLOv26 terminal evaluation |
+| M | 0.5312 | 0.5451 | Pending | Awaiting YOLOv26 terminal evaluation |
+| L | Training | Training | Pending | Awaiting both model families |
+
+The public [WALDO40 YOLOv26 repository](https://huggingface.co/MirilAI/WALDO40-YOLOv26)
+will receive checkpoints, hashes, and measured values only after validation.
+
 ## What Obnubilated changes
 
 These are representative draws from one source image. They demonstrate the