diff --git a/litellm/llms/bedrock/chat/converse_transformation.py b/litellm/llms/bedrock/chat/converse_transformation.py
index a89255a46a5..0d99bb957e0 100644
--- a/litellm/llms/bedrock/chat/converse_transformation.py
+++ b/litellm/llms/bedrock/chat/converse_transformation.py
@@ -286,6 +286,10 @@ class AmazonConverseConfig(BaseConfig):
     def _requires_min_max_tokens(model: str) -> bool:
         return re.search(r"openai\.gpt-\d|xai\.grok-", model) is not None
 
+    @staticmethod
+    def _is_openai_gpt_reasoning_model(model: str) -> bool:
+        return re.search(r"openai\.gpt-\d", model) is not None
+
     def _is_nova_2_model(self, model: str) -> bool:
         """
         Check if the model is a Nova 2 model that supports reasoningConfig.
@@ -416,12 +420,16 @@ class AmazonConverseConfig(BaseConfig):
         Handle the reasoning_effort parameter based on the model type.
 
         - GPT-OSS models: passed through unchanged via additionalModelRequestFields.
+        - OpenAI GPT-5.x and GPT-6 models: mapped to ``reasoning.effort`` via additionalModelRequestFields.
         - Nova 2 models: transformed to reasoningConfig.
         - Anthropic models: mapped to ``thinking`` (and ``output_config.effort`` on
           adaptive Claude 4.6 / 4.7).
         """
         if "gpt-oss" in model:
             optional_params["reasoning_effort"] = reasoning_effort
+        elif self._is_openai_gpt_reasoning_model(model):
+            reasoning: Final[BedrockConverseGptReasoningEffortBlock] = {"effort": reasoning_effort}
+            optional_params["reasoning"] = reasoning  # rebind-ok: out-param store like siblings
         elif self._is_nova_2_model(model):
             reasoning_config: Final = self._transform_reasoning_effort_to_reasoning_config(reasoning_effort)
             optional_params.update(reasoning_config)
@@ -553,7 +561,11 @@ class AmazonConverseConfig(BaseConfig):
             # only anthropic and mistral support tool choice config. otherwise (E.g. cohere) will fail the call - https://docs.aws.amazon.com/bedrock/latest/APIReference/API_runtime_ToolChoice.html
             supported_params.append("tool_choice")
 
-        if "gpt-oss" in model:
+        if (
+            "gpt-oss" in model
+            or self._is_openai_gpt_reasoning_model(model)
+            or self._is_openai_gpt_reasoning_model(base_model)
+        ):
             supported_params.append("reasoning_effort")
         elif self._is_nova_2_model(model):
             # Nova 2 models support reasoning_effort (transformed to reasoningConfig)
@@ -905,7 +917,7 @@ class AmazonConverseConfig(BaseConfig):
                 optional_params["_parallel_tool_use_config"] = {
                     "tool_choice": {"type": "auto", "disable_parallel_tool_use": not value}
                 }
-            if param == "thinking":
+            if param == "thinking" and not self._is_openai_gpt_reasoning_model(model):
                 if (
                     isinstance(value, dict)
                     and value.get("type") == "adaptive"
diff --git a/litellm/types/llms/bedrock.py b/litellm/types/llms/bedrock.py
index 5665aa3277a..97d068f898d 100644
--- a/litellm/types/llms/bedrock.py
+++ b/litellm/types/llms/bedrock.py
@@ -2,7 +2,7 @@ import json
 from enum import Enum
 from typing import TYPE_CHECKING, Any, Final, Literal
 
-from typing_extensions import Required, TypedDict, override
+from typing_extensions import ReadOnly, Required, TypedDict, override
 
 from .openai import ChatCompletionToolCallChunk
 
@@ -96,6 +96,10 @@ class BedrockConverseReasoningContentBlockDelta(TypedDict, total=False):
     text: str
 
 
+class BedrockConverseGptReasoningEffortBlock(TypedDict):
+    effort: ReadOnly[str]
+
+
 class GuardrailConverseTextBlock(TypedDict, total=False):
     text: str
 
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 d94c908b5f0..7d23b371620 100644
--- a/tests/test_litellm/llms/bedrock/chat/test_converse_transformation.py
+++ b/tests/test_litellm/llms/bedrock/chat/test_converse_transformation.py
@@ -288,6 +288,74 @@ def test_reasoning_with_forced_tool_choice_switches_to_auto():
     assert optional_params["tool_choice"] == {"auto": {}}
 
 
+@pytest.mark.parametrize(
+    "model",
+    [
+        "us.openai.gpt-5.6-sol",
+        "global.openai.gpt-5.6-terra",
+        "bedrock/converse/us.openai.gpt-5.6-luna",
+        "us.openai.gpt-6-astra",
+        "bedrock/converse/global.openai.gpt-6-astra",
+    ],
+)
+def test_reasoning_effort_maps_to_reasoning_effort_for_openai_gpt5_converse(model):
+    """OpenAI GPT-5.x on Bedrock Converse routes reasoning_effort to
+    ``additionalModelRequestFields.reasoning.effort`` rather than Anthropic ``thinking``."""
+    config = AmazonConverseConfig()
+
+    assert "reasoning_effort" in config.get_supported_openai_params(model=model)
+
+    optional_params = config.map_openai_params(
+        non_default_params={"reasoning_effort": "high"},
+        optional_params={},
+        model=model,
+        drop_params=False,
+    )
+
+    assert optional_params["reasoning"] == {"effort": "high"}
+    assert "thinking" not in optional_params
+    assert "reasoning_effort" not in optional_params
+
+    _, additional_request_params, _, _ = config._prepare_request_params(optional_params, model)
+    assert additional_request_params["reasoning"] == {"effort": "high"}
+    assert "thinking" not in additional_request_params
+
+
+@pytest.mark.parametrize(
+    "model",
+    [
+        "us.openai.gpt-5.6-sol",
+        "bedrock/converse/global.openai.gpt-5.6-luna",
+        "us.openai.gpt-6-astra",
+    ],
+)
+def test_openai_gpt5_converse_never_forwards_thinking(model):
+    """GPT-5.x on Converse must never send Anthropic ``thinking``/``output_config`` (Bedrock rejects them).
+
+    Regression: ``thinking`` is not advertised as supported, and even when supplied alongside
+    ``reasoning_effort`` in either order it never survives into the request."""
+    config = AmazonConverseConfig()
+
+    supported = config.get_supported_openai_params(model=model)
+    assert "thinking" not in supported
+    assert "output_config" not in supported
+
+    thinking_block = {"type": "enabled", "budget_tokens": 2048}
+    for non_default_params in (
+        {"reasoning_effort": "high", "thinking": thinking_block},
+        {"thinking": thinking_block, "reasoning_effort": "high"},
+    ):
+        optional_params = config.map_openai_params(
+            non_default_params=dict(non_default_params),
+            optional_params={},
+            model=model,
+            drop_params=False,
+        )
+        _, additional_request_params, _, _ = config._prepare_request_params(optional_params, model)
+        assert additional_request_params["reasoning"] == {"effort": "high"}
+        assert "thinking" not in additional_request_params
+
+
 @pytest.mark.parametrize(
     "model, param, value, expected_max_tokens",
     [