From 970d9f7bdb5c34548f84b1e75497205798803301 Mon Sep 17 00:00:00 2001
From: Ronald Tse <ronald.tse@ribose.com>
Date: Wed, 2 Sep 2026 17:49:33 +0200
Subject: [PATCH] docs(ml): complete the frontier axis with glm-4.7-flash
 (13.00 raw DER)

Fourth row of the family regression on SadeedDiac-25; reworded
successors to the-rest-of-the-family since 4.7-flash is a
predecessor generation, not a successor.
---
 src/pages/ml.astro | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/src/pages/ml.astro b/src/pages/ml.astro
index 8a302c7..422a1fc 100644
--- a/src/pages/ml.astro
+++ b/src/pages/ml.astro
@@ -185,10 +185,11 @@ const clientModels = [
       <p class="ledger-note">
         Frontier context, same instrument (SadeedDiac-25, the same 1,200 paragraphs under the
         same windowed evaluator): GLM-5.2 diacritizes Arabic at 2.51 DER with plain completion;
-        its successors do not — GLM-5.3-Flash at 8.57, GLM-5.3 at 9.98 raw DER — and the loss is
-        wrong vowels, not writing convention. As frontier optimization shifts agentic, dedicated
-        distilled models remain the reproducible, protocol-pinned way to ship this task. Every
-        row with its decode protocol and bootstrap CIs: <a
+        the rest of the family does not — GLM-5.3-Flash at 8.57, GLM-5.3 at 9.98, glm-4.7-flash
+        at 13.00 raw DER — and the loss is wrong vowels, not writing convention. As frontier
+        optimization shifts agentic, dedicated distilled models remain the reproducible,
+        protocol-pinned way to ship this task. Every row with its decode protocol and bootstrap
+        CIs: <a
           href="https://github.com/interscript/rababa/blob/main/docs/RESULTS.md"
           >rababa/docs/RESULTS.md</a
         >.