diff --git a/tests/e2e/run_deepseek_megatron_parallelism.sh b/tests/e2e/run_deepseek_megatron_parallelism.sh
index cb3ad6ccb86..8b8fcb182ef 100644
--- a/tests/e2e/run_deepseek_megatron_parallelism.sh
+++ b/tests/e2e/run_deepseek_megatron_parallelism.sh
@@ -21,7 +21,8 @@ python3 -m verl.trainer.main_ppo --config-path=config \
     actor_rollout_ref.actor.ppo_micro_batch_size_per_gpu=4 \
     actor_rollout_ref.actor.megatron.pipeline_model_parallel_size=2 \
     actor_rollout_ref.actor.megatron.virtual_pipeline_model_parallel_size=2 \
-    actor_rollout_ref.actor.megatron.tensor_model_parallel_size=4 \
+    actor_rollout_ref.actor.megatron.context_parallel_size=2 \
+    actor_rollout_ref.actor.megatron.tensor_model_parallel_size=2 \
     actor_rollout_ref.actor.use_kl_loss=False \
     actor_rollout_ref.rollout.log_prob_micro_batch_size_per_gpu=8 \
     actor_rollout_ref.rollout.tensor_model_parallel_size=2 \
@@ -30,6 +31,7 @@ python3 -m verl.trainer.main_ppo --config-path=config \
     actor_rollout_ref.ref.log_prob_micro_batch_size_per_gpu=16 \
     actor_rollout_ref.ref.megatron.pipeline_model_parallel_size=2 \
     actor_rollout_ref.ref.megatron.virtual_pipeline_model_parallel_size=2 \
+    actor_rollout_ref.ref.megatron.context_parallel_size=2 \
     actor_rollout_ref.ref.megatron.tensor_model_parallel_size=2 \
     critic.optim.lr=2e-5 \
     critic.model.path=$HOME/models/deepseek-ai/deepseek-coder-1.3b-instruct \
@@ -37,6 +39,7 @@ python3 -m verl.trainer.main_ppo --config-path=config \
     critic.ppo_micro_batch_size_per_gpu=4 \
     critic.megatron.pipeline_model_parallel_size=2 \
     critic.megatron.virtual_pipeline_model_parallel_size=2 \
+    critic.megatron.context_parallel_size=2 \
     critic.megatron.tensor_model_parallel_size=2 \
     algorithm.use_kl_in_reward=True \
     algorithm.kl_penalty=kl \
diff --git a/tests/e2e/run_qwen_megatron_parallelism.sh b/tests/e2e/run_qwen_megatron_parallelism.sh
index 480afe3dcda..2eb19cc4710 100644
--- a/tests/e2e/run_qwen_megatron_parallelism.sh
+++ b/tests/e2e/run_qwen_megatron_parallelism.sh
@@ -24,6 +24,7 @@ python3 -m verl.trainer.main_ppo --config-path=config \
     actor_rollout_ref.actor.ppo_micro_batch_size_per_gpu=4 \
     actor_rollout_ref.actor.megatron.pipeline_model_parallel_size=2 \
     actor_rollout_ref.actor.megatron.virtual_pipeline_model_parallel_size=2 \
+    actor_rollout_ref.actor.megatron.context_parallel_size=2 \
     actor_rollout_ref.actor.megatron.tensor_model_parallel_size=2 \
     actor_rollout_ref.actor.use_kl_loss=False \
     actor_rollout_ref.rollout.log_prob_micro_batch_size_per_gpu=8 \
@@ -33,6 +34,7 @@ python3 -m verl.trainer.main_ppo --config-path=config \
     actor_rollout_ref.ref.log_prob_micro_batch_size_per_gpu=16 \
     actor_rollout_ref.ref.megatron.pipeline_model_parallel_size=2 \
     actor_rollout_ref.ref.megatron.virtual_pipeline_model_parallel_size=2 \
+    actor_rollout_ref.ref.megatron.context_parallel_size=2 \
     actor_rollout_ref.ref.megatron.tensor_model_parallel_size=2 \
     critic.optim.lr=2e-5 \
     critic.model.path=$HOME/models/Qwen/Qwen2.5-0.5B \
@@ -40,6 +42,7 @@ python3 -m verl.trainer.main_ppo --config-path=config \
     critic.ppo_micro_batch_size_per_gpu=4 \
     critic.megatron.pipeline_model_parallel_size=2 \
     critic.megatron.virtual_pipeline_model_parallel_size=2 \
+    critic.megatron.context_parallel_size=2 \
     critic.megatron.tensor_model_parallel_size=2 \
     algorithm.use_kl_in_reward=True \
     algorithm.kl_penalty=kl \