From 5f2a695fc857b2a5432124bebdb085d2e79c25c9 Mon Sep 17 00:00:00 2001
From: johngreek <2006605+jgreek@users.noreply.github.com>
Date: Thu, 22 Jan 2026 22:13:40 -0500
Subject: [PATCH] [Fix] Anthropic models on Azure AI cache pricing (#19532)

---
 ...odel_prices_and_context_window_backup.json | 12 ++++++
 model_prices_and_context_window.json          | 12 ++++++
 .../anthropic/test_azure_ai_cache_pricing.py  | 42 +++++++++++++++++++
 3 files changed, 66 insertions(+)
 create mode 100644 tests/test_litellm/llms/anthropic/test_azure_ai_cache_pricing.py

diff --git a/litellm/model_prices_and_context_window_backup.json b/litellm/model_prices_and_context_window_backup.json
index 470d598a25f..16a345247a4 100644
--- a/litellm/model_prices_and_context_window_backup.json
+++ b/litellm/model_prices_and_context_window_backup.json
@@ -1312,6 +1312,9 @@
         "supports_function_calling": true
     },
     "azure_ai/claude-haiku-4-5": {
+        "cache_creation_input_token_cost": 1.25e-06,
+        "cache_creation_input_token_cost_above_1hr": 2e-06,
+        "cache_read_input_token_cost": 1e-07,
         "input_cost_per_token": 1e-06,
         "litellm_provider": "azure_ai",
         "max_input_tokens": 200000,
@@ -1330,6 +1333,9 @@
         "supports_vision": true
     },
     "azure_ai/claude-opus-4-5": {
+        "cache_creation_input_token_cost": 6.25e-06,
+        "cache_creation_input_token_cost_above_1hr": 1e-05,
+        "cache_read_input_token_cost": 5e-07,
         "input_cost_per_token": 5e-06,
         "litellm_provider": "azure_ai",
         "max_input_tokens": 200000,
@@ -1348,6 +1354,9 @@
         "supports_vision": true
     },
     "azure_ai/claude-opus-4-1": {
+        "cache_creation_input_token_cost": 1.875e-05,
+        "cache_creation_input_token_cost_above_1hr": 3e-05,
+        "cache_read_input_token_cost": 1.5e-06,
         "input_cost_per_token": 1.5e-05,
         "litellm_provider": "azure_ai",
         "max_input_tokens": 200000,
@@ -1366,6 +1375,9 @@
         "supports_vision": true
     },
     "azure_ai/claude-sonnet-4-5": {
+        "cache_creation_input_token_cost": 3.75e-06,
+        "cache_creation_input_token_cost_above_1hr": 6e-06,
+        "cache_read_input_token_cost": 3e-07,
         "input_cost_per_token": 3e-06,
         "litellm_provider": "azure_ai",
         "max_input_tokens": 200000,
diff --git a/model_prices_and_context_window.json b/model_prices_and_context_window.json
index 470d598a25f..16a345247a4 100644
--- a/model_prices_and_context_window.json
+++ b/model_prices_and_context_window.json
@@ -1312,6 +1312,9 @@
         "supports_function_calling": true
     },
     "azure_ai/claude-haiku-4-5": {
+        "cache_creation_input_token_cost": 1.25e-06,
+        "cache_creation_input_token_cost_above_1hr": 2e-06,
+        "cache_read_input_token_cost": 1e-07,
         "input_cost_per_token": 1e-06,
         "litellm_provider": "azure_ai",
         "max_input_tokens": 200000,
@@ -1330,6 +1333,9 @@
         "supports_vision": true
     },
     "azure_ai/claude-opus-4-5": {
+        "cache_creation_input_token_cost": 6.25e-06,
+        "cache_creation_input_token_cost_above_1hr": 1e-05,
+        "cache_read_input_token_cost": 5e-07,
         "input_cost_per_token": 5e-06,
         "litellm_provider": "azure_ai",
         "max_input_tokens": 200000,
@@ -1348,6 +1354,9 @@
         "supports_vision": true
     },
     "azure_ai/claude-opus-4-1": {
+        "cache_creation_input_token_cost": 1.875e-05,
+        "cache_creation_input_token_cost_above_1hr": 3e-05,
+        "cache_read_input_token_cost": 1.5e-06,
         "input_cost_per_token": 1.5e-05,
         "litellm_provider": "azure_ai",
         "max_input_tokens": 200000,
@@ -1366,6 +1375,9 @@
         "supports_vision": true
     },
     "azure_ai/claude-sonnet-4-5": {
+        "cache_creation_input_token_cost": 3.75e-06,
+        "cache_creation_input_token_cost_above_1hr": 6e-06,
+        "cache_read_input_token_cost": 3e-07,
         "input_cost_per_token": 3e-06,
         "litellm_provider": "azure_ai",
         "max_input_tokens": 200000,
diff --git a/tests/test_litellm/llms/anthropic/test_azure_ai_cache_pricing.py b/tests/test_litellm/llms/anthropic/test_azure_ai_cache_pricing.py
new file mode 100644
index 00000000000..45c988a21b8
--- /dev/null
+++ b/tests/test_litellm/llms/anthropic/test_azure_ai_cache_pricing.py
@@ -0,0 +1,42 @@
+"""
+Test that Azure AI Anthropic models have cache pricing configured.
+Verifies the fix for issue #19532.
+"""
+
+import sys
+import os
+
+sys.path.insert(0, os.path.abspath("../../../../../"))
+
+import litellm
+from litellm import get_model_info
+from litellm.litellm_core_utils.get_model_cost_map import get_model_cost_map
+import pytest
+
+
+@pytest.fixture(autouse=True)
+def reload_model_costs():
+    """Reload model costs from JSON before each test."""
+    litellm.model_cost = get_model_cost_map(url=None)
+    yield
+
+
+@pytest.mark.parametrize(
+    "model,expected_cache_creation_cost,expected_cache_read_cost",
+    [
+        ("claude-haiku-4-5", 1.25e-06, 1e-07),
+        ("claude-opus-4-5", 6.25e-06, 5e-07),
+        ("claude-opus-4-1", 1.875e-05, 1.5e-06),
+        ("claude-sonnet-4-5", 3.75e-06, 3e-07),
+    ],
+)
+def test_azure_ai_claude_cache_pricing(
+    model, expected_cache_creation_cost, expected_cache_read_cost
+):
+    """Test that Azure AI Claude models have correct cache pricing."""
+    model_info = get_model_info(model=model, custom_llm_provider="azure_ai")
+
+    assert model_info.get("cache_creation_input_token_cost") is not None
+    assert model_info.get("cache_read_input_token_cost") is not None
+    assert model_info.get("cache_creation_input_token_cost") == expected_cache_creation_cost
+    assert model_info.get("cache_read_input_token_cost") == expected_cache_read_cost