diff --git a/.github/workflows/vllm_ascend.yml b/.github/workflows/vllm_ascend.yml
index 7c485862e09..eaf2a210baf 100644
--- a/.github/workflows/vllm_ascend.yml
+++ b/.github/workflows/vllm_ascend.yml
@@ -110,7 +110,7 @@ jobs:
         run: |
           ray stop --force
           python3 examples/data_preprocess/gsm8k.py --local_dataset_path ${HOME}/models/hf_data/gsm8k
-      - name: Test the latest vLLM Rollout async with agent loop
+      - name: Test the latest vLLM Rollout async with agent loop with seed 0(non deter)
         run: |
           export HCCL_HOST_SOCKET_PORT_RANGE=auto
           export HCCL_NPU_SOCKET_PORT_RANGE=auto
diff --git a/tests/experimental/agent_loop/test_multi_modal.py b/tests/experimental/agent_loop/test_multi_modal.py
index 35f4600b524..556f3d30ce5 100644
--- a/tests/experimental/agent_loop/test_multi_modal.py
+++ b/tests/experimental/agent_loop/test_multi_modal.py
@@ -66,6 +66,7 @@ def init_config() -> DictConfig:
     config.actor_rollout_ref.rollout.prompt_length = 10240
     config.actor_rollout_ref.rollout.response_length = 4096
     config.actor_rollout_ref.rollout.n = 4
+    config.actor_rollout_ref.rollout.seed = 0
     config.actor_rollout_ref.rollout.agent.num_workers = 2
     config.actor_rollout_ref.rollout.skip_tokenizer_init = True