diff --git a/litellm/model_prices_and_context_window_backup.json b/litellm/model_prices_and_context_window_backup.json
index b9415c17d818..c1fe80e856a3 100644
--- a/litellm/model_prices_and_context_window_backup.json
+++ b/litellm/model_prices_and_context_window_backup.json
@@ -50346,14 +50346,13 @@
     "us.xai.grok-4.6": {
         "input_cost_per_token": 2.2e-06,
         "output_cost_per_token": 6.6e-06,
-        "cache_read_input_token_cost": 5.5e-07,
         "litellm_provider": "bedrock_converse",
         "max_input_tokens": 500000,
         "max_output_tokens": 500000,
         "max_tokens": 500000,
         "mode": "chat",
         "supports_function_calling": true,
-        "supports_prompt_caching": true,
+        "supports_prompt_caching": false,
         "supports_reasoning": true,
         "supports_tool_choice": true,
         "supports_vision": true
@@ -50361,14 +50360,13 @@
     "global.xai.grok-4.6": {
         "input_cost_per_token": 2e-06,
         "output_cost_per_token": 6e-06,
-        "cache_read_input_token_cost": 5e-07,
         "litellm_provider": "bedrock_converse",
         "max_input_tokens": 500000,
         "max_output_tokens": 500000,
         "max_tokens": 500000,
         "mode": "chat",
         "supports_function_calling": true,
-        "supports_prompt_caching": true,
+        "supports_prompt_caching": false,
         "supports_reasoning": true,
         "supports_tool_choice": true,
         "supports_vision": true
diff --git a/model_prices_and_context_window.json b/model_prices_and_context_window.json
index b9415c17d818..c1fe80e856a3 100644
--- a/model_prices_and_context_window.json
+++ b/model_prices_and_context_window.json
@@ -50346,14 +50346,13 @@
     "us.xai.grok-4.6": {
         "input_cost_per_token": 2.2e-06,
         "output_cost_per_token": 6.6e-06,
-        "cache_read_input_token_cost": 5.5e-07,
         "litellm_provider": "bedrock_converse",
         "max_input_tokens": 500000,
         "max_output_tokens": 500000,
         "max_tokens": 500000,
         "mode": "chat",
         "supports_function_calling": true,
-        "supports_prompt_caching": true,
+        "supports_prompt_caching": false,
         "supports_reasoning": true,
         "supports_tool_choice": true,
         "supports_vision": true
@@ -50361,14 +50360,13 @@
     "global.xai.grok-4.6": {
         "input_cost_per_token": 2e-06,
         "output_cost_per_token": 6e-06,
-        "cache_read_input_token_cost": 5e-07,
         "litellm_provider": "bedrock_converse",
         "max_input_tokens": 500000,
         "max_output_tokens": 500000,
         "max_tokens": 500000,
         "mode": "chat",
         "supports_function_calling": true,
-        "supports_prompt_caching": true,
+        "supports_prompt_caching": false,
         "supports_reasoning": true,
         "supports_tool_choice": true,
         "supports_vision": true
diff --git a/tests/test_litellm/integrations/test_anthropic_cache_control_hook.py b/tests/test_litellm/integrations/test_anthropic_cache_control_hook.py
index 3a736e2a8899..51634c4b723f 100644
--- a/tests/test_litellm/integrations/test_anthropic_cache_control_hook.py
+++ b/tests/test_litellm/integrations/test_anthropic_cache_control_hook.py
@@ -1611,6 +1611,18 @@ def test_model_without_caching_support_not_injected(self, monkeypatch):
         monkeypatch.setattr(litellm, "enable_anthropic_prompt_caching", True)
         assert self._points(model="anthropic.claude-3-5-sonnet-20240620-v1:0", provider="bedrock") == []
 
+    @pytest.mark.parametrize("model", ["us.xai.grok-4.6", "global.xai.grok-4.6"])
+    def test_bedrock_grok_not_injected(self, monkeypatch, local_model_cost_map, model):
+        """Bedrock does not support prompt caching for Grok. The cost-map entries must
+        carry supports_prompt_caching=false so no breakpoints are injected; otherwise
+        Bedrock rejects the whole request ("You invoked an unsupported model or your
+        request did not allow prompt caching")."""
+        from litellm.utils import supports_prompt_caching
+
+        monkeypatch.setattr(litellm, "enable_anthropic_prompt_caching", True)
+        assert supports_prompt_caching(model=model, custom_llm_provider="bedrock") is False
+        assert self._points(model=model, provider="bedrock") == []
+
     def test_stands_down_when_client_sent_cache_control(self, monkeypatch):
         monkeypatch.setattr(litellm, "enable_anthropic_prompt_caching", True)
         messages = [