From 5f0934309918b2b65c39c65f5eae3d1321e9ca3c Mon Sep 17 00:00:00 2001
From: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Date: Tue, 15 Sep 2026 20:50:01 +0000
Subject: [PATCH] test(bedrock): pin Nova cache-read prices with dated AWS
 Price List citation

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---
 ..._cross_region_inference_profile_mapping.py | 25 +++++++++++--------
 1 file changed, 14 insertions(+), 11 deletions(-)

diff --git a/tests/test_litellm/llms/bedrock/test_cross_region_inference_profile_mapping.py b/tests/test_litellm/llms/bedrock/test_cross_region_inference_profile_mapping.py
index b12ca31c694..b1e85c4f5ea 100644
--- a/tests/test_litellm/llms/bedrock/test_cross_region_inference_profile_mapping.py
+++ b/tests/test_litellm/llms/bedrock/test_cross_region_inference_profile_mapping.py
@@ -229,22 +229,25 @@ def test_bedrock_gpt_5_6_offers_tools_and_reasoning_effort_but_not_thinking(prof
     assert "output_config" not in supported
 
 
+# Cache-read prices are the `*-cache-read-input-tokens` usagetype rows of the AWS Price List API, us-east-1,
+# https://pricing.us-east-1.amazonaws.com/offers/v1.0/aws/AmazonBedrock/current/us-east-1/index.json on 2026-09-15
 @pytest.mark.parametrize(
-    "model",
+    "model,expected_cache_read",
     [
-        "amazon.nova-lite-v1:0",
-        "us.amazon.nova-lite-v1:0",
-        "amazon.nova-micro-v1:0",
-        "us.amazon.nova-micro-v1:0",
-        "amazon.nova-pro-v1:0",
-        "us.amazon.nova-pro-v1:0",
-        "us.amazon.nova-premier-v1:0",
+        ("amazon.nova-lite-v1:0", 1.5e-8),
+        ("us.amazon.nova-lite-v1:0", 1.5e-8),
+        ("amazon.nova-micro-v1:0", 8.75e-9),
+        ("us.amazon.nova-micro-v1:0", 8.75e-9),
+        ("amazon.nova-pro-v1:0", 2e-7),
+        ("us.amazon.nova-pro-v1:0", 2e-7),
+        ("us.amazon.nova-premier-v1:0", 6.25e-7),
     ],
 )
-def test_bedrock_nova_cache_read_prices(model, local_model_cost_map):
+def test_bedrock_nova_cache_read_prices(
+    model, expected_cache_read, local_model_cost_map
+):
     model_info = litellm.model_cost[model]
-    expected_cache_read = model_info["cache_read_input_token_cost"]
-    assert 0 < expected_cache_read < model_info["input_cost_per_token"]
+    assert model_info["cache_read_input_token_cost"] == expected_cache_read
     usage = Usage(
         prompt_tokens=1_000,
         completion_tokens=100,