From fc186f5613d7fd14301d5e89c19793711f786e3b Mon Sep 17 00:00:00 2001
From: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Date: Fri, 18 Sep 2026 20:02:31 +0000
Subject: [PATCH] fix(bedrock): clamp maxTokens to the 16-token minimum for
 OpenAI GPT models on Converse

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---
 .../bedrock/chat/converse_transformation.py   | 13 ++++++++++-
 .../chat/test_converse_transformation.py      | 22 +++++++++++++++++++
 2 files changed, 34 insertions(+), 1 deletion(-)

diff --git a/litellm/llms/bedrock/chat/converse_transformation.py b/litellm/llms/bedrock/chat/converse_transformation.py
index 1a4f27e2ddd..4ef42664b9f 100644
--- a/litellm/llms/bedrock/chat/converse_transformation.py
+++ b/litellm/llms/bedrock/chat/converse_transformation.py
@@ -107,6 +107,7 @@ BEDROCK_COMPUTER_USE_TOOLS: Final = [
     "bash_",
     "text_editor_",
 ]
+BEDROCK_OPENAI_GPT_MIN_MAX_TOKENS: Final = 16
 
 # Beta header patterns that are not supported by Bedrock Converse API
 # These will be filtered out to prevent errors
@@ -377,6 +378,12 @@ class AmazonConverseConfig(BaseConfig):
     def _is_openai_gpt_reasoning_model(model: str) -> bool:
         return re.search(r"openai\.gpt-\d", model) is not None
 
+    @staticmethod
+    def _enforce_min_max_tokens(max_tokens: object) -> object:
+        if isinstance(max_tokens, int) and max_tokens < BEDROCK_OPENAI_GPT_MIN_MAX_TOKENS:
+            return BEDROCK_OPENAI_GPT_MIN_MAX_TOKENS
+        return max_tokens
+
     def _is_nova_2_model(self, model: str) -> bool:
         """
         Check if the model is a Nova 2 model that supports reasoningConfig.
@@ -999,7 +1006,11 @@ class AmazonConverseConfig(BaseConfig):
                     is_thinking_enabled=is_thinking_enabled,
                 )
             if param == "max_tokens" or param == "max_completion_tokens":
-                optional_params["maxTokens"] = value
+                optional_params["maxTokens"] = (
+                    self._enforce_min_max_tokens(cast(object, value))
+                    if self._is_openai_gpt_reasoning_model(model)
+                    else value
+                )
             if param == "stream":
                 optional_params["stream"] = value
             if param == "stop":
diff --git a/tests/test_litellm/llms/bedrock/chat/test_converse_transformation.py b/tests/test_litellm/llms/bedrock/chat/test_converse_transformation.py
index 96c78c1cf75..f05b6ae97fe 100644
--- a/tests/test_litellm/llms/bedrock/chat/test_converse_transformation.py
+++ b/tests/test_litellm/llms/bedrock/chat/test_converse_transformation.py
@@ -457,6 +457,28 @@ def test_reasoning_with_forced_tool_choice_switches_to_auto():
     assert optional_params["tool_choice"] == {"auto": {}}
 
 
+@pytest.mark.parametrize(
+    "model, param, value, expected_max_tokens",
+    [
+        ("us.openai.gpt-6-astra", "max_tokens", 1, 16),
+        ("us.openai.gpt-6-astra", "max_completion_tokens", 1, 16),
+        ("us.openai.gpt-6-astra", "max_tokens", 64, 64),
+        ("anthropic.claude-sonnet-4-5-20250929-v1:0", "max_tokens", 1, 1),
+    ],
+)
+def test_map_openai_params_enforces_minimum_max_tokens_for_openai_gpt_reasoning_models(
+    model: str, param: str, value: int, expected_max_tokens: int
+):
+    optional_params = AmazonConverseConfig().map_openai_params(
+        non_default_params={param: value},
+        optional_params={},
+        model=model,
+        drop_params=False,
+    )
+
+    assert optional_params["maxTokens"] == expected_max_tokens
+
+
 @pytest.mark.parametrize(
     "model",
     [