diff --git a/automation/src/test/resources/reference.conf b/automation/src/test/resources/reference.conf
index 99b6622513..6a4ec40f1d 100644
--- a/automation/src/test/resources/reference.conf
+++ b/automation/src/test/resources/reference.conf
@@ -7,7 +7,7 @@ leonardo {
     baseImageUrl = "us.gcr.io/broad-dsp-gcr-public/terra-jupyter-base:1.1.3"
     gcrWelderUri = "us.gcr.io/broad-dsp-gcr-public/welder-server"
     dockerHubWelderUri = "broadinstitute/welder-server"
-    rstudioBioconductorImageUrl = "us.gcr.io/broad-dsp-gcr-public/anvil-rstudio-bioconductor:3.20.0"
+    rstudioBioconductorImageUrl = "us.gcr.io/broad-dsp-gcr-public/anvil-rstudio-bioconductor:3.19.0"
     //each fiab will have a unique topic name sourced from leonardo.conf, this is never published to in automation tests to ensure pub/sub components can auth properly, but never read from to avoid conflicts.
     topicName = "leonardo-pubsub-test"
     location = "us-central1-a"
diff --git a/http/src/main/resources/reference.conf b/http/src/main/resources/reference.conf
index 072b19f64a..cbc0f5044e 100644
--- a/http/src/main/resources/reference.conf
+++ b/http/src/main/resources/reference.conf
@@ -70,7 +70,7 @@ dataproc {
   }
 
   # Cached dataproc image used by Terra
-  customDataprocImage = "projects/broad-dsp-gcr-public/global/images/leo-dataproc-image-2-1-11-debian11-2024-12-06-04-19-37"
+  customDataprocImage = "projects/broad-dsp-gcr-public/global/images/leo-dataproc-image-2-1-11-debian11-2024-12-11-20-54-48"
 
   # The ratio of memory allocated to spark. 0.8 = 80%.
   # Hail/Spark users generally allocate 80% of the ram to the JVM.
@@ -111,7 +111,7 @@ dataproc {
 }
 
 gce {
-  customGceImage = "projects/broad-dsp-gcr-public/global/images/leo-gce-image-2024-12-06-03-09-27"
+  customGceImage = "projects/broad-dsp-gcr-public/global/images/leo-gce-image-2024-12-11-20-55-03"
   userDiskDeviceName = "user-disk"
   defaultScopes = [
     "https://www.googleapis.com/auth/userinfo.email",
diff --git a/jenkins/dataproc-custom-images/prepare-custom-leonardo-jupyter-dataproc-image.sh b/jenkins/dataproc-custom-images/prepare-custom-leonardo-jupyter-dataproc-image.sh
index e065028fbe..fed9354ef6 100755
--- a/jenkins/dataproc-custom-images/prepare-custom-leonardo-jupyter-dataproc-image.sh
+++ b/jenkins/dataproc-custom-images/prepare-custom-leonardo-jupyter-dataproc-image.sh
@@ -24,7 +24,7 @@ terra_jupyter_gatk="us.gcr.io/broad-dsp-gcr-public/terra-jupyter-gatk:2.3.8"
 terra_jupyter_aou="us.gcr.io/broad-dsp-gcr-public/terra-jupyter-aou:2.2.14"
 welder_server="us.gcr.io/broad-dsp-gcr-public/welder-server:8667bfe"
 openidc_proxy="broadinstitute/openidc-proxy:2.3.1_2"
-anvil_rstudio_bioconductor="us.gcr.io/broad-dsp-gcr-public/anvil-rstudio-bioconductor:3.20.0"
+anvil_rstudio_bioconductor="us.gcr.io/broad-dsp-gcr-public/anvil-rstudio-bioconductor:3.19.0"
 
 # Note that this is the version used currently by AOU in production, the one above can be staged for testing
 terra_jupyter_aou_old="us.gcr.io/broad-dsp-gcr-public/terra-jupyter-aou:2.2.13"
diff --git a/jenkins/gce-custom-images/prepare_gce_image.sh b/jenkins/gce-custom-images/prepare_gce_image.sh
index adbc9b8efb..bbd9aaa52e 100755
--- a/jenkins/gce-custom-images/prepare_gce_image.sh
+++ b/jenkins/gce-custom-images/prepare_gce_image.sh
@@ -23,7 +23,7 @@ terra_jupyter_gatk="us.gcr.io/broad-dsp-gcr-public/terra-jupyter-gatk:2.3.8"
 terra_jupyter_aou="us.gcr.io/broad-dsp-gcr-public/terra-jupyter-aou:2.2.14"
 welder_server="us.gcr.io/broad-dsp-gcr-public/welder-server:8667bfe"
 openidc_proxy="broadinstitute/openidc-proxy:2.3.1_2"
-anvil_rstudio_bioconductor="us.gcr.io/broad-dsp-gcr-public/anvil-rstudio-bioconductor:3.20.0"
+anvil_rstudio_bioconductor="us.gcr.io/broad-dsp-gcr-public/anvil-rstudio-bioconductor:3.19.0"
 
 # Note that this is the version used currently by AOU in production, the one above can be staged for testing
 terra_jupyter_aou_old="us.gcr.io/broad-dsp-gcr-public/terra-jupyter-aou:2.2.13"