Merge 069a645419 into eacb670b6e

litellm/__init__.py

@@ -685,6 +685,7 @@ qwen_ai_platform_models: Set = set()
 moonshot_models: Set = set()
 publicai_models: Set = set()
 darkbloom_models: Set = set()
+token_kiosk_models: Final[Set[str]] = set()
 openinfer_models: Final[Set[str]] = set()
 kyma_models: Final[set[str]] = set()  # mutable-ok: populated when the shared model catalog loads
 streamlake_models: Final[Set[str]] = set()
@@ -979,6 +980,8 @@ def _populate_provider_model_sets(model_cost_map: Mapping[str, Mapping[str, obje
             publicai_models.add(key)
         elif value.get("litellm_provider") == "darkbloom":
             darkbloom_models.add(key)
+        elif value.get("litellm_provider") == "token_kiosk":
+            token_kiosk_models.add(key)
         elif value.get("litellm_provider") == "openinfer":
             openinfer_models.add(key)
         elif value.get("litellm_provider") == "kyma":
@@ -1165,6 +1168,7 @@ model_list = list(
     | moonshot_models
     | publicai_models
     | darkbloom_models
+    | token_kiosk_models
     | openinfer_models
     | kyma_models
     | streamlake_models
@@ -1299,6 +1303,7 @@ def _build_models_by_provider() -> dict:
         "moonshot": moonshot_models,
         "publicai": publicai_models,
         "darkbloom": darkbloom_models,
+        "token_kiosk": token_kiosk_models,
         "openinfer": openinfer_models,
         "kyma": kyma_models,
         "streamlake": streamlake_models,

litellm/constants.py

@@ -1132,6 +1132,7 @@ openai_compatible_providers: Final[list[str]] = [
     "lm_studio",
     "galadriel",
     "zerogpu",
+    "token_kiosk",
     "github_copilot",  # GitHub Copilot Chat API
     "chatgpt",  # ChatGPT subscription API
     "novita",

litellm/llms/openai_like/providers.json

@@ -314,5 +314,14 @@
       "supports_function_calling": true,
       "supports_reasoning": true
     }
+  },
+  "token_kiosk": {
+    "base_url": "https://api-token-kiosk.gaib.ai/v1",
+    "api_key_env": "TOKEN_KIOSK_API_KEY",
+    "api_base_env": "TOKEN_KIOSK_API_BASE",
+    "provider_info": {
+      "supports_function_calling": true,
+      "supports_reasoning": true
+    }
   }
 }

litellm/types/utils.py

@@ -4785,6 +4785,7 @@ class LlmProviders(str, Enum):
     INFERENCEAPIS = "inferenceapis"
     HOSTINGER_ROUTER = "hostinger_router"
     ZEROGPU = "zerogpu"
+    TOKEN_KIOSK = "token_kiosk"
     DARKBLOOM = "darkbloom"
     META = "meta"
     SAIL = "sail"

tests/unit/llms/openai_like/test_token_kiosk_provider.py

@@ -0,0 +1,64 @@
+import json
+from typing import Final
+
+import pytest
+import respx
+
+import litellm
+from litellm.litellm_core_utils.get_llm_provider_logic import get_llm_provider
+
+
+def test_token_kiosk_resolves_environment_and_explicit_credentials(monkeypatch: pytest.MonkeyPatch) -> None:
+    monkeypatch.setenv("TOKEN_KIOSK_API_KEY", "fake-token-kiosk-key")
+    assert get_llm_provider(model="token_kiosk/claude-3-5-sonnet") == (
+        "claude-3-5-sonnet",
+        "token_kiosk",
+        "fake-token-kiosk-key",
+        "https://api-token-kiosk.gaib.ai/v1",
+    )
+    assert get_llm_provider(
+        model="token_kiosk/claude-3-5-sonnet",
+        api_key="explicit-key",
+        api_base="https://custom-router.example/v1",
+    ) == (
+        "claude-3-5-sonnet",
+        "token_kiosk",
+        "explicit-key",
+        "https://custom-router.example/v1",
+    )
+
+
+def test_token_kiosk_chat_completion_dispatch() -> None:
+    with respx.mock() as upstream:
+        route: Final = upstream.post("https://api-token-kiosk.gaib.ai/v1/chat/completions").respond(
+            200,
+            json={
+                "id": "chatcmpl-token-kiosk-test",
+                "object": "chat.completion",
+                "created": 1724000000,
+                "model": "claude-3-5-sonnet",
+                "choices": [
+                    {
+                        "index": 0,
+                        "message": {"role": "assistant", "content": "Hello from Token Kiosk!"},
+                        "finish_reason": "stop",
+                    }
+                ],
+                "usage": {
+                    "prompt_tokens": 10,
+                    "completion_tokens": 5,
+                    "total_tokens": 15,
+                },
+            },
+        )
+        response: Final = litellm.completion(
+            model="token_kiosk/claude-3-5-sonnet",
+            api_key="test-key",
+            messages=[{"role": "user", "content": "Hello"}],
+        )
+
+        assert route.called
+        assert route.calls.last.request.headers["authorization"] == "Bearer test-key"
+        body = json.loads(route.calls.last.request.content)
+        assert body["model"] == "claude-3-5-sonnet"
+        assert response.choices[0].message.content == "Hello from Token Kiosk!"