diff --git a/CLAUDE.md b/CLAUDE.md
index 0faa41d..66c38d4 100644
--- a/CLAUDE.md
+++ b/CLAUDE.md
@@ -87,3 +87,6 @@ Speaker 9B 5GB + Thinker Guard-4B 2.5GB = 7.5GB co-resident, 8GB libre. Guard-8B
 
 ## Cascade speed/qualite (24/05)
 9B+Guard4B: thinker dec11/speaker8.4, 7.5GB, ~9s/tour. 9B+Guard8B: 7.2/8.4, 9.5GB, ~11s. Les 2 FR coherent. Reco=Guard4B (marge+vitesse). Texte cote-a-cote: cli/completion instables (DSP mapping fail satur), vraie eval qualite = harness JNI. fits confirmes.
+
+## Cmp 4B vs 9B: signal partiel
+9B analyse + riche, 4B quasi-vide. Les 2 comprennent FR. Mais cli=decode HTP 0.6t/s + reasoning anglais, /no_think ignore -> eval FR propre IMPOSSIBLE en shell. Reco 9B qualite, 4B RAM. Vrai cmp = JNI (decode CPU+template). Prouve sur 4 runs.