diff --git a/.github/workflows/serverless-model-tests.yml b/.github/workflows/serverless-model-tests.yml
index 6fd4c1e..cb18ddd 100644
--- a/.github/workflows/serverless-model-tests.yml
+++ b/.github/workflows/serverless-model-tests.yml
@@ -48,7 +48,7 @@ jobs:
 
       - name: Run serverless e2e test
         id: e2e_test
-        run: python scripts/serverless_e2e_test.py --config configs/qwen/qwen3_8b.yaml --build
+        run: python scripts/serverless_e2e_test.py --config configs/gpt-oss/gpt_oss_120b.yaml --build
         env:
           RUNPOD_API_KEY: ${{ secrets.RUNPOD_API_KEY_2 }}
           DOCKERHUB_REPO: ${{ vars.DOCKERHUB_REPO || 'runpod' }}
diff --git a/configs/gemma/gemma4_31b_it.yaml b/configs/gemma/gemma4_31b_it.yaml
index eaab1c3..2966736 100644
--- a/configs/gemma/gemma4_31b_it.yaml
+++ b/configs/gemma/gemma4_31b_it.yaml
@@ -8,4 +8,5 @@ kv-cache-dtype: fp8
 enforce-eager: false
 enable-prefix-caching: true
 enable-chunked-prefill: true
+vllm-release: v2.22.5 
 speculative-config: '{"model":"RedHatAI/gemma-4-31B-it-speculator.eagle3","method":"eagle3","num_speculative_tokens":3}'
diff --git a/configs/llama/llama3.1_8b_instruct.yaml b/configs/llama/llama3.1_8b_instruct.yaml
index 76136e2..12387c6 100644
--- a/configs/llama/llama3.1_8b_instruct.yaml
+++ b/configs/llama/llama3.1_8b_instruct.yaml
@@ -6,5 +6,6 @@ trust-remote-code: true
 quantization: fp8
 kv-cache-dtype: fp8
 enforce-eager: false
+vllm-release: v2.22.5 
 enable-prefix-caching: true
 speculative-config: '{"model":"RedHatAI/Llama-3.1-8B-Instruct-speculator.eagle3","method":"eagle3","num_speculative_tokens":3}'