diff --git a/nemo/collections/llm/recipes/log/default.py b/nemo/collections/llm/recipes/log/default.py
index 4d5e9223b535..94e595bdb811 100644
--- a/nemo/collections/llm/recipes/log/default.py
+++ b/nemo/collections/llm/recipes/log/default.py
@@ -32,7 +32,6 @@ def default_log(
 ) -> Config[nl.NeMoLogger]:
     ckpt = Config(
         nl.ModelCheckpoint,
-        save_best_model=False,
         save_last=True,
         save_top_k=10,
         every_n_train_steps=200,
diff --git a/nemo/lightning/pytorch/strategies/fsdp_strategy.py b/nemo/lightning/pytorch/strategies/fsdp_strategy.py
index 2a210c9bd7f0..d34d1716e6b4 100644
--- a/nemo/lightning/pytorch/strategies/fsdp_strategy.py
+++ b/nemo/lightning/pytorch/strategies/fsdp_strategy.py
@@ -216,7 +216,7 @@ def save_checkpoint(
             and self.trainer.state.fn == TrainerFn.FITTING
             and self.ckpt_save_optimizer
         ):
-            del checkpoint["optimizer_states"]
+            checkpoint["optimizer_states"] = {}
             checkpoint['optimizer'] = get_optimizer_state_dict(self.model, self.optimizers)
             pyt_to_mcore_state_dict(checkpoint['optimizer']['state'], prefix="optimizer.state.")
 
diff --git a/nemo/lightning/pytorch/strategies/megatron_strategy.py b/nemo/lightning/pytorch/strategies/megatron_strategy.py
index 4bf8c42ece02..3a0a0368bcef 100644
--- a/nemo/lightning/pytorch/strategies/megatron_strategy.py
+++ b/nemo/lightning/pytorch/strategies/megatron_strategy.py
@@ -634,7 +634,7 @@ def save_checkpoint(
             and self.trainer.state.fn == TrainerFn.FITTING
             and self.ckpt_save_optimizer
         ):
-            del checkpoint["optimizer_states"]
+            checkpoint["optimizer_states"] = {}
             checkpoint["optimizer"] = [self.optimizer_sharded_state_dict()]
 
         self.checkpoint_io.save_checkpoint(checkpoint, filepath, storage_options=storage_options)