diff --git a/config.template.toml b/config.template.toml
index e88150cd0766..4e5b0d870a29 100644
--- a/config.template.toml
+++ b/config.template.toml
@@ -154,6 +154,11 @@ model = "gpt-4o"
 # Drop any unmapped (unsupported) params without causing an exception
 #drop_params = false
 
+# Allow litellm to modify parameters to make them compatible with providers
+# for example by inserting a default message (like 'continue') when a message is empty
+# and the provider's API would give an error otherwise
+#modify_params = true
+
 # Using the prompt caching feature if provided by the LLM and supported
 #caching_prompt = true
 
diff --git a/evaluation/benchmarks/EDA/run_infer.py b/evaluation/benchmarks/EDA/run_infer.py
index cce795e954bf..cf63f37e785c 100644
--- a/evaluation/benchmarks/EDA/run_infer.py
+++ b/evaluation/benchmarks/EDA/run_infer.py
@@ -201,7 +201,7 @@ def process_instance(
 
     llm_config = None
     if args.llm_config:
-        llm_config = get_llm_config_arg(args.llm_config)
+        llm_config = get_llm_config_arg(args.llm_config, evaluation=True)
     if llm_config is None:
         raise ValueError(f'Could not find LLM config: --llm_config {args.llm_config}')
 
diff --git a/evaluation/benchmarks/agent_bench/run_infer.py b/evaluation/benchmarks/agent_bench/run_infer.py
index 2fb7213ce845..6833402741f6 100644
--- a/evaluation/benchmarks/agent_bench/run_infer.py
+++ b/evaluation/benchmarks/agent_bench/run_infer.py
@@ -306,7 +306,7 @@ def process_instance(
 
     llm_config = None
     if args.llm_config:
-        llm_config = get_llm_config_arg(args.llm_config)
+        llm_config = get_llm_config_arg(args.llm_config, evaluation=True)
 
     if llm_config is None:
         raise ValueError(f'Could not find LLM config: --llm_config {args.llm_config}')
diff --git a/evaluation/benchmarks/aider_bench/run_infer.py b/evaluation/benchmarks/aider_bench/run_infer.py
index f7796c7696de..23aab08dc60c 100644
--- a/evaluation/benchmarks/aider_bench/run_infer.py
+++ b/evaluation/benchmarks/aider_bench/run_infer.py
@@ -278,7 +278,7 @@ def process_instance(
 
     llm_config = None
     if args.llm_config:
-        llm_config = get_llm_config_arg(args.llm_config)
+        llm_config = get_llm_config_arg(args.llm_config, evaluation=True)
 
     if llm_config is None:
         raise ValueError(f'Could not find LLM config: --llm_config {args.llm_config}')
diff --git a/evaluation/benchmarks/biocoder/run_infer.py b/evaluation/benchmarks/biocoder/run_infer.py
index f5cdd44471a8..9b973a9bae3b 100644
--- a/evaluation/benchmarks/biocoder/run_infer.py
+++ b/evaluation/benchmarks/biocoder/run_infer.py
@@ -327,7 +327,7 @@ def process_instance(
 
     llm_config = None
     if args.llm_config:
-        llm_config = get_llm_config_arg(args.llm_config)
+        llm_config = get_llm_config_arg(args.llm_config, evaluation=True)
 
     if llm_config is None:
         raise ValueError(f'Could not find LLM config: --llm_config {args.llm_config}')
diff --git a/evaluation/benchmarks/bird/run_infer.py b/evaluation/benchmarks/bird/run_infer.py
index 839284148095..b43bc5341635 100644
--- a/evaluation/benchmarks/bird/run_infer.py
+++ b/evaluation/benchmarks/bird/run_infer.py
@@ -455,7 +455,7 @@ def execute_sql(db_path, sql):
 
     llm_config = None
     if args.llm_config:
-        llm_config = get_llm_config_arg(args.llm_config)
+        llm_config = get_llm_config_arg(args.llm_config, evaluation=True)
     if llm_config is None:
         raise ValueError(f'Could not find LLM config: --llm_config {args.llm_config}')
 
diff --git a/evaluation/benchmarks/browsing_delegation/run_infer.py b/evaluation/benchmarks/browsing_delegation/run_infer.py
index 5c1ab8c062e3..a52082db1e46 100644
--- a/evaluation/benchmarks/browsing_delegation/run_infer.py
+++ b/evaluation/benchmarks/browsing_delegation/run_infer.py
@@ -141,7 +141,7 @@ def process_instance(
 
     llm_config = None
     if args.llm_config:
-        llm_config = get_llm_config_arg(args.llm_config)
+        llm_config = get_llm_config_arg(args.llm_config, evaluation=True)
 
     if llm_config is None:
         raise ValueError(f'Could not find LLM config: --llm_config {args.llm_config}')
diff --git a/evaluation/benchmarks/commit0_bench/run_infer.py b/evaluation/benchmarks/commit0_bench/run_infer.py
index ef2df020310c..2f703356c2d1 100644
--- a/evaluation/benchmarks/commit0_bench/run_infer.py
+++ b/evaluation/benchmarks/commit0_bench/run_infer.py
@@ -570,7 +570,7 @@ def commit0_setup(dataset: pd.DataFrame, repo_split: str) -> pd.DataFrame:
 
     llm_config = None
     if args.llm_config:
-        llm_config = get_llm_config_arg(args.llm_config)
+        llm_config = get_llm_config_arg(args.llm_config, evaluation=True)
         llm_config.log_completions = True
 
     if llm_config is None:
diff --git a/evaluation/benchmarks/discoverybench/run_infer.py b/evaluation/benchmarks/discoverybench/run_infer.py
index 6d8dcbd89b3c..73f4f6590381 100644
--- a/evaluation/benchmarks/discoverybench/run_infer.py
+++ b/evaluation/benchmarks/discoverybench/run_infer.py
@@ -465,7 +465,7 @@ def create_dataset(repo_location: str, split: str = 'test'):
 
     llm_config = None
     if args.llm_config:
-        llm_config = get_llm_config_arg(args.llm_config)
+        llm_config = get_llm_config_arg(args.llm_config, evaluation=True)
     if llm_config is None:
         raise ValueError(f'Could not find LLM config: --llm_config {args.llm_config}')
 
diff --git a/evaluation/benchmarks/gaia/run_infer.py b/evaluation/benchmarks/gaia/run_infer.py
index fb6d4b3db050..582f70dd04e6 100644
--- a/evaluation/benchmarks/gaia/run_infer.py
+++ b/evaluation/benchmarks/gaia/run_infer.py
@@ -237,7 +237,7 @@ def process_instance(
 
     llm_config = None
     if args.llm_config:
-        llm_config = get_llm_config_arg(args.llm_config)
+        llm_config = get_llm_config_arg(args.llm_config, evaluation=True)
     if llm_config is None:
         raise ValueError(f'Could not find LLM config: --llm_config {args.llm_config}')
 
diff --git a/evaluation/benchmarks/gorilla/run_infer.py b/evaluation/benchmarks/gorilla/run_infer.py
index 6f5b6c9d4388..22db1e5a13dd 100644
--- a/evaluation/benchmarks/gorilla/run_infer.py
+++ b/evaluation/benchmarks/gorilla/run_infer.py
@@ -145,7 +145,7 @@ def process_instance(
 
     llm_config = None
     if args.llm_config:
-        llm_config = get_llm_config_arg(args.llm_config)
+        llm_config = get_llm_config_arg(args.llm_config, evaluation=True)
     if llm_config is None:
         raise ValueError(f'Could not find LLM config: --llm_config {args.llm_config}')
 
diff --git a/evaluation/benchmarks/gpqa/run_infer.py b/evaluation/benchmarks/gpqa/run_infer.py
index de4124859991..30c12245606f 100644
--- a/evaluation/benchmarks/gpqa/run_infer.py
+++ b/evaluation/benchmarks/gpqa/run_infer.py
@@ -325,7 +325,7 @@ def process_instance(
 
     llm_config = None
     if args.llm_config:
-        llm_config = get_llm_config_arg(args.llm_config)
+        llm_config = get_llm_config_arg(args.llm_config, evaluation=True)
     if llm_config is None:
         raise ValueError(f'Could not find LLM config: --llm_config {args.llm_config}')
 
diff --git a/evaluation/benchmarks/humanevalfix/run_infer.py b/evaluation/benchmarks/humanevalfix/run_infer.py
index fff2e23730d4..f60c1696be94 100644
--- a/evaluation/benchmarks/humanevalfix/run_infer.py
+++ b/evaluation/benchmarks/humanevalfix/run_infer.py
@@ -284,7 +284,7 @@ def process_instance(
 
     llm_config = None
     if args.llm_config:
-        llm_config = get_llm_config_arg(args.llm_config)
+        llm_config = get_llm_config_arg(args.llm_config, evaluation=True)
     if llm_config is None:
         raise ValueError(f'Could not find LLM config: --llm_config {args.llm_config}')
 
diff --git a/evaluation/benchmarks/logic_reasoning/run_infer.py b/evaluation/benchmarks/logic_reasoning/run_infer.py
index 116b438b3ee9..d7e5ad868476 100644
--- a/evaluation/benchmarks/logic_reasoning/run_infer.py
+++ b/evaluation/benchmarks/logic_reasoning/run_infer.py
@@ -287,7 +287,7 @@ def process_instance(
 
     llm_config = None
     if args.llm_config:
-        llm_config = get_llm_config_arg(args.llm_config)
+        llm_config = get_llm_config_arg(args.llm_config, evaluation=True)
     if llm_config is None:
         raise ValueError(f'Could not find LLM config: --llm_config {args.llm_config}')
 
diff --git a/evaluation/benchmarks/miniwob/run_infer.py b/evaluation/benchmarks/miniwob/run_infer.py
index 95e4c935756b..e85d0fd2abd0 100644
--- a/evaluation/benchmarks/miniwob/run_infer.py
+++ b/evaluation/benchmarks/miniwob/run_infer.py
@@ -230,7 +230,7 @@ def process_instance(
 
     llm_config = None
     if args.llm_config:
-        llm_config = get_llm_config_arg(args.llm_config)
+        llm_config = get_llm_config_arg(args.llm_config, evaluation=True)
     if llm_config is None:
         raise ValueError(f'Could not find LLM config: --llm_config {args.llm_config}')
 
diff --git a/evaluation/benchmarks/mint/run_infer.py b/evaluation/benchmarks/mint/run_infer.py
index 4414e1c4625f..e27aa679f801 100644
--- a/evaluation/benchmarks/mint/run_infer.py
+++ b/evaluation/benchmarks/mint/run_infer.py
@@ -278,7 +278,7 @@ def process_instance(
 
     llm_config = None
     if args.llm_config:
-        llm_config = get_llm_config_arg(args.llm_config)
+        llm_config = get_llm_config_arg(args.llm_config, evaluation=True)
     if llm_config is None:
         raise ValueError(f'Could not find LLM config: --llm_config {args.llm_config}')
 
diff --git a/evaluation/benchmarks/ml_bench/run_infer.py b/evaluation/benchmarks/ml_bench/run_infer.py
index e97746bb7609..39eca2d6705f 100644
--- a/evaluation/benchmarks/ml_bench/run_infer.py
+++ b/evaluation/benchmarks/ml_bench/run_infer.py
@@ -291,7 +291,7 @@ def process_instance(instance: Any, metadata: EvalMetadata, reset_logger: bool =
 
     llm_config = None
     if args.llm_config:
-        llm_config = get_llm_config_arg(args.llm_config)
+        llm_config = get_llm_config_arg(args.llm_config, evaluation=True)
     if llm_config is None:
         raise ValueError(f'Could not find LLM config: --llm_config {args.llm_config}')
 
diff --git a/evaluation/benchmarks/scienceagentbench/run_infer.py b/evaluation/benchmarks/scienceagentbench/run_infer.py
index efa6c9e42c57..7e7d5ee5564c 100644
--- a/evaluation/benchmarks/scienceagentbench/run_infer.py
+++ b/evaluation/benchmarks/scienceagentbench/run_infer.py
@@ -271,7 +271,7 @@ def process_instance(
 
     llm_config = None
     if args.llm_config:
-        llm_config = get_llm_config_arg(args.llm_config)
+        llm_config = get_llm_config_arg(args.llm_config, evaluation=True)
     if llm_config is None:
         raise ValueError(f'Could not find LLM config: --llm_config {args.llm_config}')
 
diff --git a/evaluation/benchmarks/swe_bench/run_infer.py b/evaluation/benchmarks/swe_bench/run_infer.py
index 01111f75d126..8b1c36e32e13 100644
--- a/evaluation/benchmarks/swe_bench/run_infer.py
+++ b/evaluation/benchmarks/swe_bench/run_infer.py
@@ -9,7 +9,6 @@
 from datasets import load_dataset
 
 import openhands.agenthub
-
 from evaluation.utils.shared import (
     EvalException,
     EvalMetadata,
@@ -76,7 +75,7 @@ def get_instruction(instance: pd.Series, metadata: EvalMetadata):
         '4. Rerun your reproduce script and confirm that the error is fixed!\n'
         '5. Think about edgecases and make sure your fix handles them as well\n'
         "Your thinking should be thorough and so it's fine if it's very long.\n"
-        )
+    )
 
     if RUN_WITH_BROWSING:
         instruction += (
@@ -489,7 +488,7 @@ def filter_dataset(dataset: pd.DataFrame, filter_column: str) -> pd.DataFrame:
 
     llm_config = None
     if args.llm_config:
-        llm_config = get_llm_config_arg(args.llm_config)
+        llm_config = get_llm_config_arg(args.llm_config, evaluation=True)
         llm_config.log_completions = True
 
     if llm_config is None:
diff --git a/evaluation/benchmarks/toolqa/run_infer.py b/evaluation/benchmarks/toolqa/run_infer.py
index c99f15a89ae9..c730966bb77a 100644
--- a/evaluation/benchmarks/toolqa/run_infer.py
+++ b/evaluation/benchmarks/toolqa/run_infer.py
@@ -180,7 +180,7 @@ def process_instance(instance: Any, metadata: EvalMetadata, reset_logger: bool =
 
     llm_config = None
     if args.llm_config:
-        llm_config = get_llm_config_arg(args.llm_config)
+        llm_config = get_llm_config_arg(args.llm_config, evaluation=True)
     if llm_config is None:
         raise ValueError(f'Could not find LLM config: --llm_config {args.llm_config}')
 
diff --git a/evaluation/benchmarks/webarena/run_infer.py b/evaluation/benchmarks/webarena/run_infer.py
index 531f134fd988..1d2eae37f7f3 100644
--- a/evaluation/benchmarks/webarena/run_infer.py
+++ b/evaluation/benchmarks/webarena/run_infer.py
@@ -211,7 +211,7 @@ def process_instance(
 
     llm_config = None
     if args.llm_config:
-        llm_config = get_llm_config_arg(args.llm_config)
+        llm_config = get_llm_config_arg(args.llm_config, evaluation=True)
     if llm_config is None:
         raise ValueError(f'Could not find LLM config: --llm_config {args.llm_config}')
 
diff --git a/openhands/core/config/llm_config.py b/openhands/core/config/llm_config.py
index 4e60d4a281ae..8cd2b7171447 100644
--- a/openhands/core/config/llm_config.py
+++ b/openhands/core/config/llm_config.py
@@ -44,6 +44,7 @@ class LLMConfig:
         log_completions_folder: The folder to log LLM completions to. Required if log_completions is True.
         draft_editor: A more efficient LLM to use for file editing. Introduced in [PR 3985](https://github.com/All-Hands-AI/OpenHands/pull/3985).
         custom_tokenizer: A custom tokenizer to use for token counting.
+        modify_params: Allow litellm to modify parameters to make them compatible with the provider. For example, insert default messages when empty. Defaults to True.
     """
 
     model: str = 'claude-3-5-sonnet-20241022'
@@ -79,6 +80,7 @@ class LLMConfig:
     log_completions_folder: str = os.path.join(LOG_DIR, 'completions')
     draft_editor: Optional['LLMConfig'] = None
     custom_tokenizer: str | None = None
+    modify_params: bool = True
 
     def defaults_to_dict(self) -> dict:
         """Serialize fields to a dict for the frontend, including type hints, defaults, and whether it's optional."""
diff --git a/openhands/core/config/utils.py b/openhands/core/config/utils.py
index 3aedaf952353..bbfe7d308458 100644
--- a/openhands/core/config/utils.py
+++ b/openhands/core/config/utils.py
@@ -243,9 +243,9 @@ def finalize_config(cfg: AppConfig):
         )
 
 
-# Utility function for command line --group argument
+# Utility function for command line -l (--llm-config) argument
 def get_llm_config_arg(
-    llm_config_arg: str, toml_file: str = 'config.toml'
+    llm_config_arg: str, toml_file: str = 'config.toml', evaluation: bool = False
 ) -> LLMConfig | None:
     """Get a group of llm settings from the config file.
 
@@ -268,6 +268,7 @@ def get_llm_config_arg(
     Args:
         llm_config_arg: The group of llm settings to get from the config.toml file.
         toml_file: Path to the configuration file to read from. Defaults to 'config.toml'.
+        evaluation: If True, sets modify_params=False for evaluation purposes. Defaults to False.
 
     Returns:
         LLMConfig: The LLMConfig object with the settings from the config file.
@@ -296,7 +297,10 @@ def get_llm_config_arg(
 
     # update the llm config with the specified section
     if 'llm' in toml_config and llm_config_arg in toml_config['llm']:
-        return LLMConfig.from_dict(toml_config['llm'][llm_config_arg])
+        config = LLMConfig.from_dict(toml_config['llm'][llm_config_arg])
+        if evaluation:
+            config.modify_params = False
+        return config
     logger.openhands_logger.debug(f'Loading from toml failed for {llm_config_arg}')
     return None
 
diff --git a/openhands/llm/llm.py b/openhands/llm/llm.py
index d7c7309eff3f..dfa16d977ab4 100644
--- a/openhands/llm/llm.py
+++ b/openhands/llm/llm.py
@@ -142,6 +142,7 @@ def __init__(
             temperature=self.config.temperature,
             top_p=self.config.top_p,
             drop_params=self.config.drop_params,
+            modify_params=self.config.modify_params,
         )
 
         self._completion_unwrapped = self._completion
diff --git a/pyproject.toml b/pyproject.toml
index 61dae1566fbc..bc6dcb4ae20c 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -100,6 +100,7 @@ reportlab = "*"
 [tool.coverage.run]
 concurrency = ["gevent"]
 
+
 [tool.poetry.group.runtime.dependencies]
 jupyterlab = "*"
 notebook = "*"
@@ -130,6 +131,7 @@ ignore = ["D1"]
 [tool.ruff.lint.pydocstyle]
 convention = "google"
 
+
 [tool.poetry.group.evaluation.dependencies]
 streamlit = "*"
 whatthepatch = "*"