diff --git a/apps/pii/requirements-gliner.txt b/apps/pii/requirements-gliner.txt
index a3d002f3850..1fa34dd7c07 100644
--- a/apps/pii/requirements-gliner.txt
+++ b/apps/pii/requirements-gliner.txt
@@ -6,5 +6,5 @@
 # torch is pinned in the Dockerfile instead: the CPU and CUDA targets install
 # the same version from different wheel indexes.
 gliner==0.2.27
-transformers==5.3.0
-huggingface_hub==1.3.0
+transformers==5.5.0
+huggingface_hub==1.23.0
diff --git a/docker/pii.Dockerfile b/docker/pii.Dockerfile
index 205462bad3f..66a32757af7 100644
--- a/docker/pii.Dockerfile
+++ b/docker/pii.Dockerfile
@@ -87,9 +87,16 @@ RUN --mount=type=cache,target=/root/.cache/pip \
 # Bake the GLiNER weights at build time (cached layer) so startup never
 # touches the network. HF_HUB_OFFLINE makes a missing/overridden
 # PII_GLINER_MODEL fail fast at startup instead of silently downloading.
+# The predict_entities smoke assert (labels from engines.py's
+# GLINER_ENTITY_MAPPING) fails the build if a transformers/gliner bump loads
+# the model fine but silently stops returning detections.
 ENV HF_HOME=/opt/hf-cache
 ARG GLINER_MODEL=urchade/gliner_multi_pii-v1
-RUN python -c "from gliner import GLiNER; GLiNER.from_pretrained('${GLINER_MODEL}')" && \
+RUN python -c "from gliner import GLiNER; \
+model = GLiNER.from_pretrained('${GLINER_MODEL}'); \
+ents = model.predict_entities('John Smith flew to Paris on 4 May 2021.', ['person', 'location', 'date']); \
+labels = {e['label'] for e in ents}; \
+assert {'person', 'location', 'date'} <= labels, f'GLiNER smoke inference failed, got: {ents}'" && \
     chmod -R a+rX /opt/hf-cache
 ENV HF_HUB_OFFLINE=1