From e8f311429ea9afa09195bed21f078bbd50dd791e Mon Sep 17 00:00:00 2001
From: mateo-berri <277851410+mateo-berri@users.noreply.github.com>
Date: Sat, 5 Sep 2026 19:42:15 -0700
Subject: [PATCH] fix(cost-map): stop advertising reasoning_effort max on
 azure_ai/gpt-6-astra

Foundry rejects reasoning_effort max on the gpt-6-astra deployment with a 400 that
names none, low, medium, high, and xhigh as the supported values, so the card no
longer lists max. The request path never gated max (only xhigh is opt-in), so this
only changes /model_group/info and router capability gating. The azure/ twin stays
as is because it was not verified on an Azure OpenAI host
---
 .../model_prices_and_context_window_backup.json    |  2 +-
 model_prices_and_context_window.json               |  2 +-
 .../test_reasoning_effort_capability.py            | 14 +++++++++++++-
 3 files changed, 15 insertions(+), 3 deletions(-)

diff --git a/litellm/model_prices_and_context_window_backup.json b/litellm/model_prices_and_context_window_backup.json
index 5b4652d38c7..b659c3b65e5 100644
--- a/litellm/model_prices_and_context_window_backup.json
+++ b/litellm/model_prices_and_context_window_backup.json
@@ -3518,7 +3518,7 @@
         ],
         "supports_computer_use": true,
         "supports_function_calling": true,
-        "supports_max_reasoning_effort": true,
+        "supports_max_reasoning_effort": false,
         "supports_minimal_reasoning_effort": false,
         "supports_native_streaming": true,
         "supports_none_reasoning_effort": true,
diff --git a/model_prices_and_context_window.json b/model_prices_and_context_window.json
index 5b4652d38c7..b659c3b65e5 100644
--- a/model_prices_and_context_window.json
+++ b/model_prices_and_context_window.json
@@ -3518,7 +3518,7 @@
         ],
         "supports_computer_use": true,
         "supports_function_calling": true,
-        "supports_max_reasoning_effort": true,
+        "supports_max_reasoning_effort": false,
         "supports_minimal_reasoning_effort": false,
         "supports_native_streaming": true,
         "supports_none_reasoning_effort": true,
diff --git a/tests/test_litellm/router_utils/test_reasoning_effort_capability.py b/tests/test_litellm/router_utils/test_reasoning_effort_capability.py
index 3e1f26b6e1c..fa3a6dcd95a 100644
--- a/tests/test_litellm/router_utils/test_reasoning_effort_capability.py
+++ b/tests/test_litellm/router_utils/test_reasoning_effort_capability.py
@@ -394,7 +394,6 @@ class TestGpt6AstraAdvertisesItsDocumentedLevels:
         [
             ("azure/gpt-6-astra", "azure"),
             ("azure/us/gpt-6-astra", "azure"),
-            ("azure_ai/gpt-6-astra", "azure_ai"),
         ],
     )
     def test_a_foundry_deployment_also_advertises_none(self, local_model_cost_map, model, custom_llm_provider):
@@ -413,3 +412,16 @@ class TestGpt6AstraAdvertisesItsDocumentedLevels:
             "xhigh",
             "max",
         )
+
+    def test_a_foundry_azure_ai_deployment_advertises_none_but_not_max(self, local_model_cost_map):
+        from litellm.utils import _get_model_info_helper
+
+        model_info = dict(_get_model_info_helper(model="azure_ai/gpt-6-astra", custom_llm_provider="azure_ai"))
+
+        assert resolve_supported_reasoning_efforts(model_info, deployment_is_mapped=True) == (
+            "none",
+            "low",
+            "medium",
+            "high",
+            "xhigh",
+        )