Restore published comparison charts with provenance notes

README.md

@@ -33,6 +33,18 @@ matching, metadata-language rules, calibration and swarm-availability estimates.
 Its classifier reads names, file lists and supplied counters. It runs on a CPU;
 no GPU or paid inference API is required.
 
+## Comparison charts
+
+**Latest published figures — provenance under review.** These archived charts retain their original model labels and scores. Their displayed values and vector-based source records still need reconciliation; they do not establish verified current-release performance. These Ultra hybrid comparisons do not measure standalone INT8 accuracy or payload safety. [Evaluation scope and source records](https://github.com/spencernorton/armiger/blob/main/docs/evaluation.md).
+
+![Latest published Armiger Ultra comparison with cloud teacher models; provenance under review](https://huggingface.co/NorviTech/armiger-ultra/resolve/efd514648bd86c964e7ec837a2689c682c5d56bd/assets/armiger_vs_llms_overall_and_6dims.png)
+
+Cloud teacher comparisons use separate common-support and swarm cohorts; panel-specific sample counts and metrics must remain separate.
+
+![Latest published Armiger Ultra comparison with open-source torrent classifiers and parsers; provenance under review](https://huggingface.co/NorviTech/armiger-ultra/resolve/efd514648bd86c964e7ec837a2689c682c5d56bd/assets/armiger_vs_torrent_classifiers_6dims.png)
+
+Classifier and parser comparisons cover separate category, catalogue, language, metadata-screening and swarm tasks. Original PNGs and recorded results are unchanged.
+
 ## Resources and performance
 
 **ONNX file sizes are not working RAM.** The standalone ONNX interface, GitHub
@@ -101,8 +113,11 @@ a connector are a separate operation.
 
 Published comparison graphics are under provenance review: the displayed
 3.11.0 / 2,484-item overview and its 3.0.0 / 2,835-item data receipt do not
-describe the same results. They are withdrawn from this card rather than
-presented as verified current-release comparisons. Historical hybrid results
+describe the same results. The latest published comparison charts are restored
+above as archival figures with that provenance limit. Detailed charts also
+combine newer aggregate values with older vector-based statistical panels.
+Their restored visibility does not establish current-release qualification.
+Historical hybrid results
 also do not establish the standalone neural path's accuracy. See
 [evaluation status](https://github.com/spencernorton/armiger/blob/main/docs/evaluation.md).
 No new balanced benchmark or universal accuracy, zero-leakage or safety rate is