From c9c9a4f3a241a9d03eeb96befea8a7a7737ae8fe Mon Sep 17 00:00:00 2001
From: shivam <shivam@berri.ai>
Date: Fri, 24 Jul 2026 00:28:04 +0000
Subject: [PATCH] fix(pricing): correct gpt-5.4-mini/nano max_input_tokens to
 272000

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---
 ...odel_prices_and_context_window_backup.json | 16 ++---
 model_prices_and_context_window.json          | 24 +++----
 .../test_gpt_5_4_model_metadata.py            | 66 +++++++++++++++++++
 3 files changed, 86 insertions(+), 20 deletions(-)
 create mode 100644 tests/test_litellm/test_gpt_5_4_model_metadata.py

diff --git a/litellm/model_prices_and_context_window_backup.json b/litellm/model_prices_and_context_window_backup.json
index d3917886060..fa152cc28aa 100644
--- a/litellm/model_prices_and_context_window_backup.json
+++ b/litellm/model_prices_and_context_window_backup.json
@@ -3214,7 +3214,7 @@
         "input_cost_per_token_priority": 1.5e-06,
         "input_cost_per_token_above_272k_tokens_priority": 3e-06,
         "litellm_provider": "azure_ai",
-        "max_input_tokens": 400000,
+        "max_input_tokens": 272000,
         "max_output_tokens": 128000,
         "max_tokens": 128000,
         "mode": "chat",
@@ -3260,7 +3260,7 @@
         "input_cost_per_token_priority": 1.5e-06,
         "input_cost_per_token_above_272k_tokens_priority": 3e-06,
         "litellm_provider": "azure_ai",
-        "max_input_tokens": 400000,
+        "max_input_tokens": 272000,
         "max_output_tokens": 128000,
         "max_tokens": 128000,
         "mode": "chat",
@@ -3306,7 +3306,7 @@
         "input_cost_per_token_priority": 4e-07,
         "input_cost_per_token_above_272k_tokens_priority": 8e-07,
         "litellm_provider": "azure_ai",
-        "max_input_tokens": 400000,
+        "max_input_tokens": 272000,
         "max_output_tokens": 128000,
         "max_tokens": 128000,
         "mode": "chat",
@@ -3352,7 +3352,7 @@
         "input_cost_per_token_priority": 4e-07,
         "input_cost_per_token_above_272k_tokens_priority": 8e-07,
         "litellm_provider": "azure_ai",
-        "max_input_tokens": 400000,
+        "max_input_tokens": 272000,
         "max_output_tokens": 128000,
         "max_tokens": 128000,
         "mode": "chat",
@@ -6953,7 +6953,7 @@
         "cache_read_input_token_cost": 7.5e-08,
         "input_cost_per_token": 7.5e-07,
         "litellm_provider": "azure",
-        "max_input_tokens": 1050000,
+        "max_input_tokens": 272000,
         "max_output_tokens": 128000,
         "max_tokens": 128000,
         "mode": "chat",
@@ -6988,7 +6988,7 @@
         "cache_read_input_token_cost": 7.5e-08,
         "input_cost_per_token": 7.5e-07,
         "litellm_provider": "azure",
-        "max_input_tokens": 1050000,
+        "max_input_tokens": 272000,
         "max_output_tokens": 128000,
         "max_tokens": 128000,
         "mode": "chat",
@@ -7023,7 +7023,7 @@
         "cache_read_input_token_cost": 2e-08,
         "input_cost_per_token": 2e-07,
         "litellm_provider": "azure",
-        "max_input_tokens": 1050000,
+        "max_input_tokens": 272000,
         "max_output_tokens": 128000,
         "max_tokens": 128000,
         "mode": "chat",
@@ -7058,7 +7058,7 @@
         "cache_read_input_token_cost": 2e-08,
         "input_cost_per_token": 2e-07,
         "litellm_provider": "azure",
-        "max_input_tokens": 1050000,
+        "max_input_tokens": 272000,
         "max_output_tokens": 128000,
         "max_tokens": 128000,
         "mode": "chat",
diff --git a/model_prices_and_context_window.json b/model_prices_and_context_window.json
index c9d871fc41d..4d728c71919 100644
--- a/model_prices_and_context_window.json
+++ b/model_prices_and_context_window.json
@@ -3214,7 +3214,7 @@
         "input_cost_per_token_priority": 1.5e-06,
         "input_cost_per_token_above_272k_tokens_priority": 3e-06,
         "litellm_provider": "azure_ai",
-        "max_input_tokens": 400000,
+        "max_input_tokens": 272000,
         "max_output_tokens": 128000,
         "max_tokens": 128000,
         "mode": "chat",
@@ -3260,7 +3260,7 @@
         "input_cost_per_token_priority": 1.5e-06,
         "input_cost_per_token_above_272k_tokens_priority": 3e-06,
         "litellm_provider": "azure_ai",
-        "max_input_tokens": 400000,
+        "max_input_tokens": 272000,
         "max_output_tokens": 128000,
         "max_tokens": 128000,
         "mode": "chat",
@@ -3306,7 +3306,7 @@
         "input_cost_per_token_priority": 4e-07,
         "input_cost_per_token_above_272k_tokens_priority": 8e-07,
         "litellm_provider": "azure_ai",
-        "max_input_tokens": 400000,
+        "max_input_tokens": 272000,
         "max_output_tokens": 128000,
         "max_tokens": 128000,
         "mode": "chat",
@@ -3352,7 +3352,7 @@
         "input_cost_per_token_priority": 4e-07,
         "input_cost_per_token_above_272k_tokens_priority": 8e-07,
         "litellm_provider": "azure_ai",
-        "max_input_tokens": 400000,
+        "max_input_tokens": 272000,
         "max_output_tokens": 128000,
         "max_tokens": 128000,
         "mode": "chat",
@@ -6953,7 +6953,7 @@
         "cache_read_input_token_cost": 7.5e-08,
         "input_cost_per_token": 7.5e-07,
         "litellm_provider": "azure",
-        "max_input_tokens": 1050000,
+        "max_input_tokens": 272000,
         "max_output_tokens": 128000,
         "max_tokens": 128000,
         "mode": "chat",
@@ -6988,7 +6988,7 @@
         "cache_read_input_token_cost": 7.5e-08,
         "input_cost_per_token": 7.5e-07,
         "litellm_provider": "azure",
-        "max_input_tokens": 1050000,
+        "max_input_tokens": 272000,
         "max_output_tokens": 128000,
         "max_tokens": 128000,
         "mode": "chat",
@@ -7023,7 +7023,7 @@
         "cache_read_input_token_cost": 2e-08,
         "input_cost_per_token": 2e-07,
         "litellm_provider": "azure",
-        "max_input_tokens": 1050000,
+        "max_input_tokens": 272000,
         "max_output_tokens": 128000,
         "max_tokens": 128000,
         "mode": "chat",
@@ -7058,7 +7058,7 @@
         "cache_read_input_token_cost": 2e-08,
         "input_cost_per_token": 2e-07,
         "litellm_provider": "azure",
-        "max_input_tokens": 1050000,
+        "max_input_tokens": 272000,
         "max_output_tokens": 128000,
         "max_tokens": 128000,
         "mode": "chat",
@@ -24028,7 +24028,7 @@
         "input_cost_per_token_batches": 3.75e-07,
         "input_cost_per_token_priority": 1.5e-06,
         "litellm_provider": "openai",
-        "max_input_tokens": 1050000,
+        "max_input_tokens": 272000,
         "max_output_tokens": 128000,
         "max_tokens": 128000,
         "mode": "chat",
@@ -24074,7 +24074,7 @@
         "input_cost_per_token_batches": 3.75e-07,
         "input_cost_per_token_priority": 1.5e-06,
         "litellm_provider": "openai",
-        "max_input_tokens": 1050000,
+        "max_input_tokens": 272000,
         "max_output_tokens": 128000,
         "max_tokens": 128000,
         "mode": "chat",
@@ -24118,7 +24118,7 @@
         "input_cost_per_token_flex": 1e-07,
         "input_cost_per_token_batches": 1e-07,
         "litellm_provider": "openai",
-        "max_input_tokens": 1050000,
+        "max_input_tokens": 272000,
         "max_output_tokens": 128000,
         "max_tokens": 128000,
         "mode": "chat",
@@ -24161,7 +24161,7 @@
         "input_cost_per_token_flex": 1e-07,
         "input_cost_per_token_batches": 1e-07,
         "litellm_provider": "openai",
-        "max_input_tokens": 1050000,
+        "max_input_tokens": 272000,
         "max_output_tokens": 128000,
         "max_tokens": 128000,
         "mode": "chat",
diff --git a/tests/test_litellm/test_gpt_5_4_model_metadata.py b/tests/test_litellm/test_gpt_5_4_model_metadata.py
new file mode 100644
index 00000000000..388cb4b0a45
--- /dev/null
+++ b/tests/test_litellm/test_gpt_5_4_model_metadata.py
@@ -0,0 +1,66 @@
+import json
+from pathlib import Path
+
+import pytest
+
+REPO_ROOT = Path(__file__).parents[2]
+MAIN_PATH = REPO_ROOT / "model_prices_and_context_window.json"
+BACKUP_PATH = REPO_ROOT / "litellm" / "model_prices_and_context_window_backup.json"
+
+MINI_NANO_MODELS = (
+    "gpt-5.4-mini",
+    "gpt-5.4-mini-2026-03-17",
+    "gpt-5.4-nano",
+    "gpt-5.4-nano-2026-03-17",
+    "azure/gpt-5.4-mini",
+    "azure/gpt-5.4-mini-2026-03-17",
+    "azure/gpt-5.4-nano",
+    "azure/gpt-5.4-nano-2026-03-17",
+    "azure_ai/gpt-5.4-mini",
+    "azure_ai/gpt-5.4-mini-2026-03-17",
+    "azure_ai/gpt-5.4-nano",
+    "azure_ai/gpt-5.4-nano-2026-03-17",
+)
+
+LONG_CONTEXT_MODELS = (
+    "gpt-5.4",
+    "gpt-5.4-pro",
+    "azure/gpt-5.4",
+    "azure_ai/gpt-5.4",
+)
+
+
+def _load(path):
+    with open(path) as f:
+        return json.load(f)
+
+
+@pytest.mark.parametrize("model", MINI_NANO_MODELS)
+def test_gpt_5_4_mini_nano_max_input_tokens(model):
+    """gpt-5.4-mini/nano share the 400k context window, so max input is 272k.
+
+    Regression for azure/azure_ai variants that had leaked the base gpt-5.4
+    1,050,000 window (and the raw 400,000 context window) into max_input_tokens.
+    """
+    info = _load(MAIN_PATH).get(model)
+    assert info is not None, f"{model} missing from model_prices_and_context_window.json"
+    assert info["max_input_tokens"] == 272000
+    assert info["max_output_tokens"] == 128000
+    assert info["max_tokens"] == 128000
+
+
+@pytest.mark.parametrize("model", LONG_CONTEXT_MODELS)
+def test_gpt_5_4_base_keeps_long_context_window(model):
+    """Base gpt-5.4 and gpt-5.4-pro keep the full 1,050,000 token input window."""
+    info = _load(MAIN_PATH).get(model)
+    assert info is not None, f"{model} missing from model_prices_and_context_window.json"
+    assert info["max_input_tokens"] == 1050000
+    assert info["max_output_tokens"] == 128000
+
+
+@pytest.mark.parametrize("model", MINI_NANO_MODELS + LONG_CONTEXT_MODELS)
+def test_gpt_5_4_backup_matches_main(model):
+    """The bundled cost map must stay in sync with the canonical file."""
+    assert _load(BACKUP_PATH).get(model) == _load(MAIN_PATH).get(
+        model
+    ), f"{model} differs between main and backup model cost maps"