Merge 97ae9ad7d3 into 9dbfb060bd

litellm/model_prices_and_context_window_backup.json

@@ -27666,6 +27666,21 @@
         "supports_response_schema": true,
         "supports_vision": true
     },
+    "chatgpt/gpt-5.4-mini": {
+        "litellm_provider": "chatgpt",
+        "max_input_tokens": 272000,
+        "max_output_tokens": 128000,
+        "max_tokens": 128000,
+        "mode": "responses",
+        "supported_endpoints": [
+            "/v1/chat/completions",
+            "/v1/responses"
+        ],
+        "supports_function_calling": true,
+        "supports_parallel_function_calling": true,
+        "supports_response_schema": true,
+        "supports_vision": true
+    },
     "chatgpt/gpt-5.4-pro": {
         "litellm_provider": "chatgpt",
         "max_input_tokens": 1050000,

model_prices_and_context_window.json

@@ -27666,6 +27666,21 @@
         "supports_response_schema": true,
         "supports_vision": true
     },
+    "chatgpt/gpt-5.4-mini": {
+        "litellm_provider": "chatgpt",
+        "max_input_tokens": 272000,
+        "max_output_tokens": 128000,
+        "max_tokens": 128000,
+        "mode": "responses",
+        "supported_endpoints": [
+            "/v1/chat/completions",
+            "/v1/responses"
+        ],
+        "supports_function_calling": true,
+        "supports_parallel_function_calling": true,
+        "supports_response_schema": true,
+        "supports_vision": true
+    },
     "chatgpt/gpt-5.4-pro": {
         "litellm_provider": "chatgpt",
         "max_input_tokens": 1050000,

tests/test_litellm/test_chatgpt_model_config.py

@@ -0,0 +1,62 @@
+import json
+import os
+from typing import TypeAlias, cast
+
+import pytest
+
+from litellm.main import responses_api_bridge_check
+
+JsonObject: TypeAlias = dict[str, object]
+
+REPO_ROOT = os.path.join(os.path.dirname(__file__), "../..")
+MODEL_NAME = "chatgpt/gpt-5.4-mini"
+
+
+def _load_cost_map(path: str) -> dict[str, JsonObject]:
+    with open(path) as f:
+        return cast(dict[str, JsonObject], json.load(f))
+
+
+def _load_root_cost_map() -> dict[str, JsonObject]:
+    return _load_cost_map(os.path.join(REPO_ROOT, "model_prices_and_context_window.json"))
+
+
+def _load_backup_cost_map() -> dict[str, JsonObject]:
+    return _load_cost_map(os.path.join(REPO_ROOT, "litellm/model_prices_and_context_window_backup.json"))
+
+
+@pytest.mark.parametrize(
+    "cost_map",
+    [_load_root_cost_map(), _load_backup_cost_map()],
+    ids=["root", "bundled_backup"],
+)
+def test_chatgpt_gpt_5_4_mini_is_responses_model(cost_map: dict[str, JsonObject]) -> None:
+    info = cost_map[MODEL_NAME]
+
+    assert info["litellm_provider"] == "chatgpt"
+    assert info["mode"] == "responses"
+    assert info["max_input_tokens"] == 272000
+    assert info["max_output_tokens"] == 128000
+    assert info["max_tokens"] == 128000
+    assert info["supported_endpoints"] == ["/v1/chat/completions", "/v1/responses"]
+    assert info["supports_function_calling"] is True
+    assert info["supports_parallel_function_calling"] is True
+    assert info["supports_response_schema"] is True
+    assert info["supports_vision"] is True
+
+
+def test_chatgpt_gpt_5_4_mini_bridges_chat_completions_to_responses(monkeypatch: pytest.MonkeyPatch) -> None:
+    def fake_get_model_info_helper(model: str, custom_llm_provider: str) -> JsonObject:
+        assert model == MODEL_NAME
+        assert custom_llm_provider == "chatgpt"
+        return _load_backup_cost_map()[MODEL_NAME]
+
+    monkeypatch.setattr("litellm.main._get_model_info_helper", fake_get_model_info_helper)
+
+    model_info, model = cast(
+        tuple[JsonObject, str],
+        responses_api_bridge_check(model=MODEL_NAME, custom_llm_provider="chatgpt"),
+    )
+
+    assert model == MODEL_NAME
+    assert model_info["mode"] == "responses"