diff --git a/.github/ISSUE_TEMPLATE/new-release.md b/.github/ISSUE_TEMPLATE/new-release.md
index 6ed3df8c..ceca9f5f 100644
--- a/.github/ISSUE_TEMPLATE/new-release.md
+++ b/.github/ISSUE_TEMPLATE/new-release.md
@@ -34,10 +34,10 @@ This document defines the process for releasing Gateway API Inference Extension.
    export RC=1
    ```
 
-4. The vLLM image tag defaults to `v0.7.1` for a release. Optionally, change the vLLM image tag. For example:
+4. The vLLM image tag defaults to `0.7.2` for a release. Optionally, change the vLLM image tag. For example:
 
    ```shell
-   export VLLM=0.7.2
+   export VLLM=0.7.3
    ```
 
 ## Release Process
@@ -114,7 +114,8 @@ This document defines the process for releasing Gateway API Inference Extension.
 
 9. Pushing the tag triggers Prow to build and publish the container image to the [staging registry][].
 10. Submit a PR against [k8s.io][] to add the staging image tag and SHA to [`k8s-staging-gateway-api-inference-extension/images.yaml`][yaml]. This will
-   promote the image to the production registry. **Note:** Add a link to this issue when the PR is merged.
+    promote the image to the production registry, e.g. `registry.k8s.io/gateway-api-inference-extension/epp:v${MAJOR}.${MINOR}.0`.
+    **Note:** Add a link to this issue when the PR is merged.
 11. Test the steps in the tagged quickstart guide after the PR merges, for example: `https://github.com/kubernetes-sigs/gateway-api-inference-extension/blob/v0.1.0-rc.1/pkg/README.md`.
 12. Create a [new release][]:
     1. Choose the tag that you created for the release.
diff --git a/hack/release-quickstart.sh b/hack/release-quickstart.sh
index b156b160..f4701508 100755
--- a/hack/release-quickstart.sh
+++ b/hack/release-quickstart.sh
@@ -15,8 +15,8 @@ else
   RELEASE_TAG="v${MAJOR}.${MINOR}.0-rc.${RC}"
 fi
 
-# vLLM image version (default to 0.7.1 if not defined)
-VLLM="${VLLM:-0.7.1}"
+# vLLM image version (default to 0.7.2 if not defined)
+VLLM="${VLLM:-0.7.2}"
 
 echo "Using release tag: ${RELEASE_TAG}"
 echo "Using vLLM image version: ${VLLM}"
@@ -41,12 +41,15 @@ sed -i.bak "s|kubectl apply -k https://github.com/kubernetes-sigs/gateway-api-in
 EXT_PROC="pkg/manifests/ext_proc.yaml"
 echo "Updating ${EXT_PROC} ..."
 
-# Update any image reference for the EPP container.
-# For images from registry.k8s.io:
-sed -i.bak -E "s|(registry\.k8s\.io/gateway-api-inference-extension/epp:)[^\"[:space:]]+|\1${RELEASE_TAG}|g" "$EXT_PROC"
-# In case there is still any reference from us-central1-docker.pkg.dev:
+# Update the EPP container tag.
 sed -i.bak -E "s|(us-central1-docker\.pkg\.dev/k8s-staging-images/gateway-api-inference-extension/epp:)[^\"[:space:]]+|\1${RELEASE_TAG}|g" "$EXT_PROC"
 
+# Update the EPP container image pull policy.
+sed -i.bak '/us-central1-docker.pkg.dev\/k8s-staging-images\/gateway-api-inference-extension\/epp/ { n; s/Always/IfNotPresent/ }' "$EXT_PROC"
+
+# Update the EPP container registry.
+sed -i.bak -E "s|us-central1-docker\.pkg\.dev/k8s-staging-images|registry.k8s.io|g" "$EXT_PROC"
+
 # -----------------------------------------------------------------------------
 # Update pkg/manifests/vllm/deployment.yaml
 # -----------------------------------------------------------------------------
@@ -54,10 +57,10 @@ VLLM_DEPLOY="pkg/manifests/vllm/deployment.yaml"
 echo "Updating ${VLLM_DEPLOY} ..."
 
 # Update the vLLM image version
-sed -i.bak -E "s|(vllm/vllm-openai:)[^\"[:space:]]+|\1${VLLM}|g" "$VLLM_DEPLOY"
+sed -i.bak -E "s|(vllm/vllm-openai:)[^\"[:space:]]+|\1v${VLLM}|g" "$VLLM_DEPLOY"
 
 # Also change the imagePullPolicy from Always to IfNotPresent on lines containing the vLLM image.
-sed -i.bak "/vllm\/vllm-openai/ s/Always/IfNotPresent/g" "$VLLM_DEPLOY"
+sed -i.bak '/vllm\/vllm-openai/ { n; s/Always/IfNotPresent/ }' "$VLLM_DEPLOY"
 
 # -----------------------------------------------------------------------------
 # Stage the changes
diff --git a/pkg/manifests/ext_proc.yaml b/pkg/manifests/ext_proc.yaml
index 4e82779e..a7dc7678 100644
--- a/pkg/manifests/ext_proc.yaml
+++ b/pkg/manifests/ext_proc.yaml
@@ -72,6 +72,7 @@ spec:
       containers:
       - name: inference-gateway-ext-proc
         image: us-central1-docker.pkg.dev/k8s-staging-images/gateway-api-inference-extension/epp:main
+        imagePullPolicy: Always
         args:
         - -poolName
         - "vllm-llama2-7b-pool"