diff --git a/examples/configs/recipes/llm/grpo-llama3.1-8b-instruct-2n8g-fsdp2tp1-noncolocated.yaml b/examples/configs/recipes/llm/grpo-llama3.1-8b-instruct-2n8g-fsdp2tp1-noncolocated.yaml
new file mode 100644
index 0000000000..bb8a06a0a0
--- /dev/null
+++ b/examples/configs/recipes/llm/grpo-llama3.1-8b-instruct-2n8g-fsdp2tp1-noncolocated.yaml
@@ -0,0 +1,130 @@
+grpo:
+  num_prompts_per_step: 64
+  num_generations_per_prompt: 32
+  max_rollout_turns: 1
+  max_num_steps: 500
+  normalize_rewards: true
+  use_leave_one_out_baseline: true
+  val_period: 10
+  val_at_start: false
+  max_val_samples: 256
+  val_batch_size: 256
+  seed: 42
+loss_fn:
+  reference_policy_kl_penalty: 0.01
+  ratio_clip_min: 0.2
+  ratio_clip_max: 0.2
+  ratio_clip_c: null
+  use_on_policy_kl_approximation: false
+  use_importance_sampling_correction: false
+  token_level_loss: true
+checkpointing:
+  enabled: true
+  checkpoint_dir: results/grpo-llama3.1-8b-instruct-2n8g-fsdp2tp1-noncolocated
+  metric_name: val_reward
+  higher_is_better: true
+  keep_top_k: 3
+  save_period: 10
+  checkpoint_must_save_by: null
+policy:
+  model_name: meta-llama/Llama-3.1-8B-Instruct
+  tokenizer:
+    name: meta-llama/Llama-3.1-8B-Instruct
+  train_global_batch_size: 512
+  train_micro_batch_size: 1
+  generation_batch_size: 32
+  logprob_batch_size: 2
+  max_total_sequence_length: 4096
+  precision: bfloat16
+  dtensor_cfg:
+    enabled: true
+    cpu_offload: false
+    sequence_parallel: false
+    activation_checkpointing: false
+    tensor_parallel_size: 1
+    context_parallel_size: 1
+    custom_parallel_plan: null
+  dynamic_batching:
+    enabled: True
+    train_mb_tokens: ${mul:${policy.max_total_sequence_length}, ${policy.train_micro_batch_size}}
+    logprob_mb_tokens: ${mul:${policy.max_total_sequence_length}, ${policy.logprob_batch_size}}
+    sequence_length_round: 64
+  sequence_packing:
+    enabled: false
+    train_mb_tokens: ${mul:${policy.max_total_sequence_length}, ${policy.train_micro_batch_size}}
+    logprob_mb_tokens: ${mul:${policy.max_total_sequence_length}, ${policy.logprob_batch_size}}
+    algorithm: "modified_first_fit_decreasing"
+    sequence_length_round: 64
+  make_sequence_length_divisible_by: 1
+  max_grad_norm: 1
+  optimizer:
+    name: torch.optim.AdamW
+    kwargs:
+      lr: 3e-07
+      weight_decay: 0.01
+      betas:
+        - 0.9
+        - 0.999
+      eps: 1e-08
+      foreach: false
+      fused: false
+  scheduler:
+    - name: torch.optim.lr_scheduler.LinearLR
+      kwargs:
+        start_factor: 0.1
+        end_factor: 1
+        total_iters: 13
+    - name: torch.optim.lr_scheduler.ConstantLR
+      kwargs:
+        factor: 1
+        total_iters: 10000000000
+    - milestones:
+        - 13
+  generation:
+    backend: vllm
+    max_new_tokens: 4096
+    temperature: 1
+    top_p: 1
+    top_k: null
+    stop_token_ids:
+      - 128009
+    stop_strings: null
+    vllm_cfg:
+      async_engine: true
+      precision: ${policy.precision}
+      tensor_parallel_size: 1
+      pipeline_parallel_size: 1
+      gpu_memory_utilization: 0.6
+      max_model_len: 4096
+      enforce_eager: False
+    colocated:
+      enabled: false
+      resources:
+        gpus_per_node: null
+        num_nodes: 1
+data:
+  max_input_seq_length: 4096
+  prompt_file: examples/prompts/cot.txt
+  system_prompt_file: null
+  dataset_name: OpenMathInstruct-2
+  shuffle: true
+env:
+  math:
+    num_workers: 8
+logger:
+  log_dir: logs/grpo-llama3.1-8b-instruct-2n8g-fsdp2tp1-noncolocated
+  num_val_samples_to_print: 0
+  wandb_enabled: true
+  tensorboard_enabled: true
+  mlflow_enabled: false
+  monitor_gpus: true
+  wandb:
+    project: nemo-rl
+    name: grpo-llama3.1-8b-instruct-2n8g-fsdp2tp1-noncolocated
+  tensorboard: {}
+  gpu_monitoring:
+    collection_interval: 10
+    flush_interval: 10
+cluster:
+  gpus_per_node: 8
+  num_nodes: 2
diff --git a/tests/test_suites/llm/dpo-llama3.1-8b-instruct-4n8g-fsdp2tp2-quick.v2.sh b/tests/test_suites/llm/dpo-llama3.1-8b-instruct-4n8g-fsdp2tp2-quick.v2.sh
index 0162bd8bb9..a8d2d04adc 100755
--- a/tests/test_suites/llm/dpo-llama3.1-8b-instruct-4n8g-fsdp2tp2-quick.v2.sh
+++ b/tests/test_suites/llm/dpo-llama3.1-8b-instruct-4n8g-fsdp2tp2-quick.v2.sh
@@ -40,4 +40,4 @@ if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | ma
         'data["train/preference_loss"]["1"] < 0.69316' \
         'data["train/preference_loss"]["20"] < 0.6' \
         'mean(data["timing/train/total_step_time"], -10, -1) < 7.8'
-fi 
+fi
diff --git a/tests/test_suites/llm/dpo-llama3.1-8b-instruct-4n8g-fsdp2tp4.sh b/tests/test_suites/llm/dpo-llama3.1-8b-instruct-4n8g-fsdp2tp4.sh
index df74127ba2..fbda6865f5 100755
--- a/tests/test_suites/llm/dpo-llama3.1-8b-instruct-4n8g-fsdp2tp4.sh
+++ b/tests/test_suites/llm/dpo-llama3.1-8b-instruct-4n8g-fsdp2tp4.sh
@@ -40,4 +40,4 @@ if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | ma
         'data["train/preference_loss"]["1"] < 0.69316' \
         'data["train/preference_loss"]["150"] < 0.4' \
         'mean(data["timing/train/total_step_time"], -11, -1) < 24'
-fi 
+fi
diff --git a/tests/test_suites/llm/dpo-llama3.1-8b-instruct-4n8g-megatron.v2.sh b/tests/test_suites/llm/dpo-llama3.1-8b-instruct-4n8g-megatron.v2.sh
index 8701d63d1f..7cc74e26df 100755
--- a/tests/test_suites/llm/dpo-llama3.1-8b-instruct-4n8g-megatron.v2.sh
+++ b/tests/test_suites/llm/dpo-llama3.1-8b-instruct-4n8g-megatron.v2.sh
@@ -40,4 +40,4 @@ if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | ma
         'data["train/preference_loss"]["1"] < 0.69316' \
         'data["train/preference_loss"]["150"] < 0.4' \
         'mean(data["timing/train/total_step_time"], -11, -1) < 11.5'
-fi 
+fi
diff --git a/tests/test_suites/llm/dpo-llama3.1-8b-instruct-4n8g-megatrontp2pp2-quick.sh b/tests/test_suites/llm/dpo-llama3.1-8b-instruct-4n8g-megatrontp2pp2-quick.sh
index 0bc8e13e28..497e0b8f68 100755
--- a/tests/test_suites/llm/dpo-llama3.1-8b-instruct-4n8g-megatrontp2pp2-quick.sh
+++ b/tests/test_suites/llm/dpo-llama3.1-8b-instruct-4n8g-megatrontp2pp2-quick.sh
@@ -40,4 +40,4 @@ if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | ma
         'data["train/preference_loss"]["1"] < 0.69316' \
         'data["train/preference_loss"]["20"] < 0.6' \
         'mean(data["timing/train/total_step_time"], -10) < 6.7'
-fi 
+fi
diff --git a/tests/test_suites/llm/dpo-llama3.1-8b-tulu3-1n8g-fsdp2tp1.sh b/tests/test_suites/llm/dpo-llama3.1-8b-tulu3-1n8g-fsdp2tp1.sh
index 35da749ac7..a6beabb886 100755
--- a/tests/test_suites/llm/dpo-llama3.1-8b-tulu3-1n8g-fsdp2tp1.sh
+++ b/tests/test_suites/llm/dpo-llama3.1-8b-tulu3-1n8g-fsdp2tp1.sh
@@ -40,4 +40,3 @@ if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | ma
         'data["train/preference_loss"]["1"] < 0.6932' \
         'data["train/preference_loss"]["150"] < 0.68'
 fi
-
diff --git a/tests/test_suites/llm/dpo-llama3.2-1b-instruct-1n8g-fsdp2tp1.v2.sh b/tests/test_suites/llm/dpo-llama3.2-1b-instruct-1n8g-fsdp2tp1.v2.sh
index 48691c0df4..0b0c67b312 100755
--- a/tests/test_suites/llm/dpo-llama3.2-1b-instruct-1n8g-fsdp2tp1.v2.sh
+++ b/tests/test_suites/llm/dpo-llama3.2-1b-instruct-1n8g-fsdp2tp1.v2.sh
@@ -38,4 +38,4 @@ if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | ma
         'data["train/loss"]["1"] < 0.69316' \
         'data["train/loss"]["150"] < 0.55' \
         'mean(data["timing/train/total_step_time"], -11, -1) < 1.3'
-fi 
+fi
diff --git a/tests/test_suites/llm/grpo-deepscaler-1.5b-16K.sh b/tests/test_suites/llm/grpo-deepscaler-1.5b-16K.sh
index 68c4f0c6b2..633b0d8297 100755
--- a/tests/test_suites/llm/grpo-deepscaler-1.5b-16K.sh
+++ b/tests/test_suites/llm/grpo-deepscaler-1.5b-16K.sh
@@ -66,4 +66,3 @@ cat ${RUN_LOG}.aime-16k       | grep "score=" | sed 's/.*score=\([^ ]*\).*/{"sco
 # 240 step checkpoint 0.3
 uv run tests/check_metrics.py ${RUN_LOG}-16k-metric.json \
   'data["score"] >= 0.2396'
-
diff --git a/tests/test_suites/llm/grpo-deepscaler-1.5b-24K.sh b/tests/test_suites/llm/grpo-deepscaler-1.5b-24K.sh
index b045607570..92dc48f719 100755
--- a/tests/test_suites/llm/grpo-deepscaler-1.5b-24K.sh
+++ b/tests/test_suites/llm/grpo-deepscaler-1.5b-24K.sh
@@ -65,4 +65,3 @@ cat ${RUN_LOG}.aime-24k       | grep "score=" | sed 's/.*score=\([^ ]*\).*/{"sco
  
 uv run tests/check_metrics.py ${RUN_LOG}-24k-metric.json \
   'data["score"] >= 0.2396'
-
diff --git a/tests/test_suites/llm/grpo-gemma3-27b-it-16n8g-fsdp2tp8sp-actckpt-long.sh b/tests/test_suites/llm/grpo-gemma3-27b-it-16n8g-fsdp2tp8sp-actckpt-long.sh
index 69c9899ccd..a6ce1800d9 100755
--- a/tests/test_suites/llm/grpo-gemma3-27b-it-16n8g-fsdp2tp8sp-actckpt-long.sh
+++ b/tests/test_suites/llm/grpo-gemma3-27b-it-16n8g-fsdp2tp8sp-actckpt-long.sh
@@ -37,4 +37,3 @@ if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | ma
         'mean(data["train/token_mult_prob_error"]) < 1.1' \
         'data["train/token_mult_prob_error"]["20"] < 1.1'
 fi
-
diff --git a/tests/test_suites/llm/grpo-llama3.1-8b-instruct-1n8g-megatron-fp8.sh b/tests/test_suites/llm/grpo-llama3.1-8b-instruct-1n8g-megatron-fp8.sh
index e0b2f928da..31c8e4739e 100755
--- a/tests/test_suites/llm/grpo-llama3.1-8b-instruct-1n8g-megatron-fp8.sh
+++ b/tests/test_suites/llm/grpo-llama3.1-8b-instruct-1n8g-megatron-fp8.sh
@@ -37,4 +37,3 @@ if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | ma
         'mean(data["train/token_mult_prob_error"]) < 1.1' \
         'data["train/token_mult_prob_error"]["100"] < 1.1'
 fi
-
diff --git a/tests/test_suites/llm/grpo-llama3.1-8b-instruct-2n8g-fsdp2tp1-noncolocated.sh b/tests/test_suites/llm/grpo-llama3.1-8b-instruct-2n8g-fsdp2tp1-noncolocated.sh
new file mode 100755
index 0000000000..d018032576
--- /dev/null
+++ b/tests/test_suites/llm/grpo-llama3.1-8b-instruct-2n8g-fsdp2tp1-noncolocated.sh
@@ -0,0 +1,39 @@
+#!/bin/bash
+SCRIPT_DIR=$( cd -- "$( dirname -- "${BASH_SOURCE[0]}" )" &> /dev/null && pwd)
+source $SCRIPT_DIR/common.env
+
+# ===== BEGIN CONFIG =====
+NUM_NODES=2
+STEPS_PER_RUN=30
+MAX_STEPS=30
+NUM_RUNS=$(( (MAX_STEPS + STEPS_PER_RUN - 1) / STEPS_PER_RUN ))  # Round up
+NUM_MINUTES=120
+# ===== END CONFIG =====
+
+exit_if_max_steps_reached
+
+# Run the experiment
+cd $PROJECT_ROOT
+uv run examples/run_grpo_math.py \
+    --config $CONFIG_PATH \
+    grpo.max_num_steps=$MAX_STEPS \
+    logger.log_dir=$LOG_DIR \
+    logger.wandb_enabled=True \
+    logger.wandb.project=nemo-rl \
+    logger.wandb.name=$EXP_NAME \
+    logger.monitor_gpus=True \
+    logger.tensorboard_enabled=True \
+    checkpointing.enabled=True \
+    checkpointing.checkpoint_dir=$CKPT_DIR \
+    $@ \
+    2>&1 | tee $RUN_LOG
+
+# Convert tensorboard logs to json
+uv run tests/json_dump_tb_logs.py $LOG_DIR --output_path $JSON_METRICS
+
+# Only run metrics if the target step is reached
+if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | map(tonumber) | max' $JSON_METRICS) -ge $MAX_STEPS ]]; then
+    uv run tests/check_metrics.py $JSON_METRICS \
+        'mean(data["train/token_mult_prob_error"]) < 1.1' \
+        'data["train/token_mult_prob_error"]["30"] < 1.1'
+fi
diff --git a/tests/test_suites/llm/grpo-llama3.1-8b-instruct-4n8g-fsdp2tp1-long.v3.sh b/tests/test_suites/llm/grpo-llama3.1-8b-instruct-4n8g-fsdp2tp1-long.v3.sh
index 6e64876058..af44d060cb 100755
--- a/tests/test_suites/llm/grpo-llama3.1-8b-instruct-4n8g-fsdp2tp1-long.v3.sh
+++ b/tests/test_suites/llm/grpo-llama3.1-8b-instruct-4n8g-fsdp2tp1-long.v3.sh
@@ -37,4 +37,3 @@ if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | ma
         'mean(data["train/token_mult_prob_error"]) < 1.1' \
         'data["train/token_mult_prob_error"]["100"] < 1.1'
 fi
-
diff --git a/tests/test_suites/llm/grpo-llama3.2-1b-instruct-1n8g-fsdp2tp1.v3.sh b/tests/test_suites/llm/grpo-llama3.2-1b-instruct-1n8g-fsdp2tp1.v3.sh
index 3661370fa6..562ff730e7 100755
--- a/tests/test_suites/llm/grpo-llama3.2-1b-instruct-1n8g-fsdp2tp1.v3.sh
+++ b/tests/test_suites/llm/grpo-llama3.2-1b-instruct-1n8g-fsdp2tp1.v3.sh
@@ -38,4 +38,3 @@ if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | ma
         'data["train/token_mult_prob_error"]["500"] < 1.1' \
         'mean(data["timing/train/total_step_time"], -6, -1) < 10'
 fi
-
diff --git a/tests/test_suites/llm/grpo-llama3.2-1b-instruct-1n8g-megatron.sh b/tests/test_suites/llm/grpo-llama3.2-1b-instruct-1n8g-megatron.sh
index 83071c70e3..90e309e128 100755
--- a/tests/test_suites/llm/grpo-llama3.2-1b-instruct-1n8g-megatron.sh
+++ b/tests/test_suites/llm/grpo-llama3.2-1b-instruct-1n8g-megatron.sh
@@ -38,5 +38,4 @@ if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | ma
         'data["train/token_mult_prob_error"]["500"] < 1.1' \
         'data["train/reward"]["500"] > 0.1' \
         'mean(data["timing/train/total_step_time"], -6, -1) < 10.5'
-
 fi
diff --git a/tests/test_suites/llm/grpo-qwen2.5-32b-32n8g-fsdp2tp8sp-actckpt-long.v3.sh b/tests/test_suites/llm/grpo-qwen2.5-32b-32n8g-fsdp2tp8sp-actckpt-long.v3.sh
index 7d1fdc2858..fa7fbd5bd6 100755
--- a/tests/test_suites/llm/grpo-qwen2.5-32b-32n8g-fsdp2tp8sp-actckpt-long.v3.sh
+++ b/tests/test_suites/llm/grpo-qwen2.5-32b-32n8g-fsdp2tp8sp-actckpt-long.v3.sh
@@ -37,4 +37,3 @@ if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | ma
         'mean(data["train/token_mult_prob_error"]) < 1.1' \
         'data["train/token_mult_prob_error"]["20"] < 1.1'
 fi
-
diff --git a/tests/test_suites/llm/grpo-qwen2.5-32b-32n8g-fsdp2tp8sp-actckpt.v3.sh b/tests/test_suites/llm/grpo-qwen2.5-32b-32n8g-fsdp2tp8sp-actckpt.v3.sh
index e96f3de7fa..98591ba9b3 100755
--- a/tests/test_suites/llm/grpo-qwen2.5-32b-32n8g-fsdp2tp8sp-actckpt.v3.sh
+++ b/tests/test_suites/llm/grpo-qwen2.5-32b-32n8g-fsdp2tp8sp-actckpt.v3.sh
@@ -37,4 +37,3 @@ if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | ma
         'mean(data["train/token_mult_prob_error"]) < 1.1' \
         'data["train/token_mult_prob_error"]["2"] < 1.1'
 fi
-
diff --git a/tests/test_suites/llm/grpo-qwen2.5-7b-instruct-4n8g-fsdp2tp4sp.v3.sh b/tests/test_suites/llm/grpo-qwen2.5-7b-instruct-4n8g-fsdp2tp4sp.v3.sh
index b3071fb58e..ec613562f2 100755
--- a/tests/test_suites/llm/grpo-qwen2.5-7b-instruct-4n8g-fsdp2tp4sp.v3.sh
+++ b/tests/test_suites/llm/grpo-qwen2.5-7b-instruct-4n8g-fsdp2tp4sp.v3.sh
@@ -37,4 +37,3 @@ if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | ma
         'mean(data["train/token_mult_prob_error"]) < 1.1' \
         'data["train/token_mult_prob_error"]["30"] < 1.1'
 fi
-
diff --git a/tests/test_suites/llm/grpo-qwen2.5-math-1.5b-instruct-1n8g-fsdp2tp1.v3.sh b/tests/test_suites/llm/grpo-qwen2.5-math-1.5b-instruct-1n8g-fsdp2tp1.v3.sh
index 0a31e74590..35810c4eec 100755
--- a/tests/test_suites/llm/grpo-qwen2.5-math-1.5b-instruct-1n8g-fsdp2tp1.v3.sh
+++ b/tests/test_suites/llm/grpo-qwen2.5-math-1.5b-instruct-1n8g-fsdp2tp1.v3.sh
@@ -38,4 +38,3 @@ if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | ma
         'data["train/token_mult_prob_error"]["450"] < 1.1' \
         'mean(data["timing/train/total_step_time"], 2) < 25'
 fi
-
diff --git a/tests/test_suites/llm/sft-llama3.1-8b-1n8g-fsdp2tp1-dynamicbatch.sh b/tests/test_suites/llm/sft-llama3.1-8b-1n8g-fsdp2tp1-dynamicbatch.sh
index 76c600c648..3395f54f43 100755
--- a/tests/test_suites/llm/sft-llama3.1-8b-1n8g-fsdp2tp1-dynamicbatch.sh
+++ b/tests/test_suites/llm/sft-llama3.1-8b-1n8g-fsdp2tp1-dynamicbatch.sh
@@ -35,7 +35,7 @@ uv run tests/json_dump_tb_logs.py $LOG_DIR --output_path $JSON_METRICS
 # Only run metrics if the target step is reached
 if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | map(tonumber) | max' $JSON_METRICS) -ge $MAX_STEPS ]]; then
     uv run tests/check_metrics.py $JSON_METRICS \
-	'data["train/loss"]["1"] < 0.6' \
+	    'data["train/loss"]["1"] < 0.6' \
         'data["train/loss"]["250"] < 0.36' \
         'max(data["ray/node.0.gpu.0.mem_gb"]) < 70' \
         'mean(data["timing/train/total_step_time"], 2) < 10'
diff --git a/tests/test_suites/llm/sft-llama3.1-8b-1n8g-fsdp2tp1-long.sh b/tests/test_suites/llm/sft-llama3.1-8b-1n8g-fsdp2tp1-long.sh
index 90fd03467c..4b243e8fe9 100755
--- a/tests/test_suites/llm/sft-llama3.1-8b-1n8g-fsdp2tp1-long.sh
+++ b/tests/test_suites/llm/sft-llama3.1-8b-1n8g-fsdp2tp1-long.sh
@@ -37,6 +37,6 @@ if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | ma
     uv run tests/check_metrics.py $JSON_METRICS \
         'data["train/loss"]["1"] < 0.6' \
         'data["train/loss"]["250"] < 0.36' \
-	'max(data["ray/node.0.gpu.0.mem_gb"]) < 80' \
+	    'max(data["ray/node.0.gpu.0.mem_gb"]) < 80' \
         'mean(data["timing/train/total_step_time"], 2) < 22'
 fi
diff --git a/tests/test_suites/llm/sft-llama3.1-8b-1n8g-fsdp2tp2sp.sh b/tests/test_suites/llm/sft-llama3.1-8b-1n8g-fsdp2tp2sp.sh
index 8f69d0f0b8..87ca1e9dad 100755
--- a/tests/test_suites/llm/sft-llama3.1-8b-1n8g-fsdp2tp2sp.sh
+++ b/tests/test_suites/llm/sft-llama3.1-8b-1n8g-fsdp2tp2sp.sh
@@ -40,4 +40,4 @@ if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | ma
         'data["train/loss"]["50"] < 0.38' \
         'max(data["ray/node.0.gpu.0.mem_gb"]) < 70' \
         'mean(data["timing/train/total_step_time"], 2) < 32'
-fi 
+fi
diff --git a/tests/test_suites/llm/sft-llama3.1-8b-1n8g-megatron-seqpack.sh b/tests/test_suites/llm/sft-llama3.1-8b-1n8g-megatron-seqpack.sh
index fe54af1fbd..e063b39861 100755
--- a/tests/test_suites/llm/sft-llama3.1-8b-1n8g-megatron-seqpack.sh
+++ b/tests/test_suites/llm/sft-llama3.1-8b-1n8g-megatron-seqpack.sh
@@ -36,4 +36,4 @@ if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | ma
         'data["train/loss"]["1"] < 0.6' \
         'data["train/loss"]["250"] < 0.36' \
         'mean(data["timing/train/total_step_time"], 2) < 6'
-fi 
+fi
diff --git a/tests/test_suites/llm/sft-llama3.1-8b-1n8g-megatron.sh b/tests/test_suites/llm/sft-llama3.1-8b-1n8g-megatron.sh
index bc5eae73a2..8ef0dfafe6 100755
--- a/tests/test_suites/llm/sft-llama3.1-8b-1n8g-megatron.sh
+++ b/tests/test_suites/llm/sft-llama3.1-8b-1n8g-megatron.sh
@@ -36,4 +36,4 @@ if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | ma
         'data["train/loss"]["1"] < 0.6' \
         'data["train/loss"]["250"] < 0.36' \
         'mean(data["timing/train/total_step_time"], 2) < 20'
-fi 
+fi
diff --git a/tests/test_suites/llm/sft-llama3.2-1b-1n8g-fsdp2tp1.v3.sh b/tests/test_suites/llm/sft-llama3.2-1b-1n8g-fsdp2tp1.v3.sh
index a4b44bd1f1..05305ab3f7 100755
--- a/tests/test_suites/llm/sft-llama3.2-1b-1n8g-fsdp2tp1.v3.sh
+++ b/tests/test_suites/llm/sft-llama3.2-1b-1n8g-fsdp2tp1.v3.sh
@@ -39,4 +39,3 @@ if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | ma
         'max(data["ray/node.0.gpu.0.mem_gb"]) < 25' \
         'mean(data["timing/train/total_step_time"], -6, -1) < 0.6'
 fi
-
diff --git a/tests/test_suites/llm/sft-qwen2.5-32b-4n8g-fsdp2tp8sp-actckpt.v3.sh b/tests/test_suites/llm/sft-qwen2.5-32b-4n8g-fsdp2tp8sp-actckpt.v3.sh
index d16a3d8d98..89edc6613e 100755
--- a/tests/test_suites/llm/sft-qwen2.5-32b-4n8g-fsdp2tp8sp-actckpt.v3.sh
+++ b/tests/test_suites/llm/sft-qwen2.5-32b-4n8g-fsdp2tp8sp-actckpt.v3.sh
@@ -40,4 +40,4 @@ if [[ $(jq 'to_entries | .[] | select(.key == "train/loss") | .value | keys | ma
         'data["train/loss"]["1"] < 0.37' \
         'data["train/loss"]["20"] < 0.3' \
         'max(data["ray/node.0.gpu.0.mem_gb"]) < 35'
-fi 
+fi
diff --git a/tests/test_suites/nightly.txt b/tests/test_suites/nightly.txt
index e6611fa8b5..f3841fef04 100644
--- a/tests/test_suites/nightly.txt
+++ b/tests/test_suites/nightly.txt
@@ -37,6 +37,9 @@ tests/test_suites/llm/grpo-math-qwen3-30ba3b-megatron-tp4-32k.sh
 # FP8
 tests/test_suites/llm/grpo-llama3.1-8b-instruct-1n8g-megatron-fp8.sh
 
+# Non-colocated
+tests/test_suites/llm/grpo-llama3.1-8b-instruct-2n8g-fsdp2tp1-noncolocated.sh
+
 #######
 # SFT #
 #######