diff --git a/litellm/constants.py b/litellm/constants.py index d1daf2ee2d1..25ee6c74465 100644 --- a/litellm/constants.py +++ b/litellm/constants.py @@ -1060,6 +1060,7 @@ openai_compatible_endpoints: Final[list[str]] = [ "https://api.scx.ai/v1", "https://api.prisminference.com/v1", "https://api.reka.ai/v1", + "https://api.voltagegpu.com/v1", "https://gigachat.devices.sberbank.ru/api/v1", "https://api.openinfer.ai/v1", ] @@ -1141,6 +1142,7 @@ openai_compatible_providers: Final[list[str]] = [ "prism", "sail", "reka", + "voltagegpu", ] OPENAI_AUDIO_TRANSCRIPTION_PROVIDERS: Final = frozenset({"openai"} | frozenset(openai_compatible_providers)) diff --git a/litellm/llms/openai_like/providers.json b/litellm/llms/openai_like/providers.json index 36337fddbdc..e770b400ad9 100644 --- a/litellm/llms/openai_like/providers.json +++ b/litellm/llms/openai_like/providers.json @@ -249,5 +249,13 @@ "/v1/responses", "/v1/messages" ] + }, + "voltagegpu": { + "base_url": "https://api.voltagegpu.com/v1", + "api_key_env": "VOLTAGEGPU_API_KEY", + "api_base_env": "VOLTAGEGPU_API_BASE", + "param_mappings": { + "max_completion_tokens": "max_tokens" + } } } diff --git a/litellm/provider_endpoints_support_backup.json b/litellm/provider_endpoints_support_backup.json index f95b553c2b4..0278d4fcc47 100644 --- a/litellm/provider_endpoints_support_backup.json +++ b/litellm/provider_endpoints_support_backup.json @@ -2407,6 +2407,24 @@ "interactions": true } }, + "voltagegpu": { + "display_name": "VoltageGPU (`voltagegpu`)", + "url": "https://docs.litellm.ai/docs/providers/voltagegpu", + "endpoints": { + "chat_completions": true, + "messages": true, + "responses": true, + "embeddings": false, + "image_generations": false, + "audio_transcriptions": false, + "audio_speech": false, + "moderations": false, + "batches": false, + "rerank": false, + "a2a": false, + "interactions": false + } + }, "voyage": { "display_name": "VoyageAI by MongoDB (`voyage`)", "url": "https://docs.litellm.ai/docs/providers/voyage", diff --git a/litellm/proxy/public_endpoints/provider_create_fields.json b/litellm/proxy/public_endpoints/provider_create_fields.json index 19ee465c7dc..1dcd786e0a1 100644 --- a/litellm/proxy/public_endpoints/provider_create_fields.json +++ b/litellm/proxy/public_endpoints/provider_create_fields.json @@ -3873,6 +3873,34 @@ ], "default_model_placeholder": "volcengine/" }, + { + "provider": "VOLTAGEGPU", + "provider_display_name": "VoltageGPU", + "litellm_provider": "voltagegpu", + "credential_fields": [ + { + "key": "api_base", + "label": "API Base", + "placeholder": "https://api.voltagegpu.com/v1", + "tooltip": null, + "required": false, + "field_type": "text", + "options": null, + "default_value": null + }, + { + "key": "api_key", + "label": "API Key", + "placeholder": null, + "tooltip": null, + "required": true, + "field_type": "password", + "options": null, + "default_value": null + } + ], + "default_model_placeholder": "voltagegpu/Qwen/Qwen3.8-27B-TEE" + }, { "provider": "Voyage", "provider_display_name": "VoyageAI by MongoDB", diff --git a/litellm/types/utils.py b/litellm/types/utils.py index cc7ef851111..381161ffd1b 100644 --- a/litellm/types/utils.py +++ b/litellm/types/utils.py @@ -4586,6 +4586,7 @@ class LlmProviders(str, Enum): META = "meta" SAIL = "sail" REKA = "reka" + VOLTAGEGPU = "voltagegpu" LITELLM_AGENT = "litellm_agent" CURSOR = "cursor" BEDROCK_MANTLE = "bedrock_mantle" diff --git a/provider_endpoints_support.json b/provider_endpoints_support.json index d65da3f41fd..018df4f78db 100644 --- a/provider_endpoints_support.json +++ b/provider_endpoints_support.json @@ -2778,6 +2778,24 @@ "video_generations": true } }, + "voltagegpu": { + "display_name": "VoltageGPU (`voltagegpu`)", + "url": "https://docs.litellm.ai/docs/providers/voltagegpu", + "endpoints": { + "chat_completions": true, + "messages": true, + "responses": true, + "embeddings": false, + "image_generations": false, + "audio_transcriptions": false, + "audio_speech": false, + "moderations": false, + "batches": false, + "rerank": false, + "a2a": false, + "interactions": false + } + }, "voyage": { "display_name": "VoyageAI by MongoDB (`voyage`)", "url": "https://docs.litellm.ai/docs/providers/voyage", diff --git a/tests/unit/llms/openai_like/test_voltagegpu_provider.py b/tests/unit/llms/openai_like/test_voltagegpu_provider.py new file mode 100644 index 00000000000..bd56906ba8d --- /dev/null +++ b/tests/unit/llms/openai_like/test_voltagegpu_provider.py @@ -0,0 +1,144 @@ +import json +from pathlib import Path +from typing import Final + +import pytest +import respx + +import litellm + +VOLTAGEGPU_API_BASE: Final = "https://api.voltagegpu.com/v1" +VOLTAGEGPU_MODEL: Final = "Qwen/Qwen3.8-27B-TEE" + + +def test_voltagegpu_json_registry(): + from litellm.llms.openai_like.json_loader import JSONProviderRegistry + + assert JSONProviderRegistry.exists("voltagegpu") + config = JSONProviderRegistry.get("voltagegpu") + assert config is not None + assert config.base_url == VOLTAGEGPU_API_BASE + assert config.api_key_env == "VOLTAGEGPU_API_KEY" + assert config.api_base_env == "VOLTAGEGPU_API_BASE" + assert config.param_mappings.get("max_completion_tokens") == "max_tokens" + + +def test_voltagegpu_provider_resolution(monkeypatch: pytest.MonkeyPatch): + """Model ids on VoltageGPU contain a `/` (org/model-TEE); only the first segment is the provider.""" + from litellm.litellm_core_utils.get_llm_provider_logic import get_llm_provider + + monkeypatch.setenv("VOLTAGEGPU_API_KEY", "voltagegpu-test-key") + + model, provider, api_key, api_base = get_llm_provider( + model=f"voltagegpu/{VOLTAGEGPU_MODEL}", + custom_llm_provider=None, + api_base=None, + api_key=None, + ) + + assert model == VOLTAGEGPU_MODEL + assert provider == "voltagegpu" + assert api_key == "voltagegpu-test-key" + assert api_base == VOLTAGEGPU_API_BASE + + +def test_voltagegpu_provider_keeps_explicit_credentials(monkeypatch: pytest.MonkeyPatch): + from litellm.litellm_core_utils.get_llm_provider_logic import get_llm_provider + + monkeypatch.setenv("VOLTAGEGPU_API_KEY", "voltagegpu-env-key") + + _, provider, api_key, api_base = get_llm_provider( + model=f"voltagegpu/{VOLTAGEGPU_MODEL}", + custom_llm_provider=None, + api_base="https://voltagegpu.internal.example/v1", + api_key="voltagegpu-explicit-key", + ) + + assert provider == "voltagegpu" + assert api_key == "voltagegpu-explicit-key" + assert api_base == "https://voltagegpu.internal.example/v1" + + +def test_voltagegpu_max_completion_tokens_is_mapped_to_max_tokens(): + """VoltageGPU documents `max_tokens` only, so `max_completion_tokens` is renamed on the way out.""" + from litellm.llms.openai_like.dynamic_config import create_config_class + from litellm.llms.openai_like.json_loader import JSONProviderRegistry + + config = create_config_class(JSONProviderRegistry.get("voltagegpu"))() + + result = config.map_openai_params( + {"max_completion_tokens": 64, "temperature": 0.2}, {}, VOLTAGEGPU_MODEL, False + ) + + assert result["max_tokens"] == 64 + assert "max_completion_tokens" not in result + assert result["temperature"] == 0.2 + + +def test_voltagegpu_is_available_in_add_model_form(): + fields_path = Path(litellm.__file__).parent / "proxy" / "public_endpoints" / "provider_create_fields.json" + providers = json.loads(fields_path.read_text()) + voltagegpu = next(provider for provider in providers if provider["litellm_provider"] == "voltagegpu") + + assert voltagegpu["provider"] == "VOLTAGEGPU" + assert voltagegpu["provider_display_name"] == "VoltageGPU" + assert voltagegpu["default_model_placeholder"] == f"voltagegpu/{VOLTAGEGPU_MODEL}" + assert {field["key"]: field["required"] for field in voltagegpu["credential_fields"]} == { + "api_base": False, + "api_key": True, + } + + +def test_voltagegpu_supported_endpoints(): + matrix_path = Path(litellm.__file__).parent / "provider_endpoints_support_backup.json" + providers = json.loads(matrix_path.read_text())["providers"] + + assert providers["voltagegpu"]["endpoints"] == { + "chat_completions": True, + "messages": True, + "responses": True, + "embeddings": False, + "image_generations": False, + "audio_transcriptions": False, + "audio_speech": False, + "moderations": False, + "batches": False, + "rerank": False, + "a2a": False, + "interactions": False, + } + + +def test_voltagegpu_chat_completion_request(): + with respx.mock() as upstream: + route: Final = upstream.post(f"{VOLTAGEGPU_API_BASE}/chat/completions").respond( + 200, + json={ + "id": "chatcmpl_voltagegpu", + "object": "chat.completion", + "created": 1_792_752_740, + "model": VOLTAGEGPU_MODEL, + "choices": [ + { + "index": 0, + "message": {"role": "assistant", "content": "Hello from VoltageGPU"}, + "finish_reason": "stop", + } + ], + "usage": {"prompt_tokens": 4, "completion_tokens": 4, "total_tokens": 8}, + }, + ) + response: Final = litellm.completion( + model=f"voltagegpu/{VOLTAGEGPU_MODEL}", + messages=[{"role": "user", "content": "Say hello"}], + api_key="voltagegpu-test-key", + ) + + request: Final = route.calls.last.request + body: Final = json.loads(request.content) + assert route.call_count == 1 + assert str(request.url) == f"{VOLTAGEGPU_API_BASE}/chat/completions" + assert request.headers["authorization"] == "Bearer voltagegpu-test-key" + assert body["model"] == VOLTAGEGPU_MODEL + assert body["messages"] == [{"role": "user", "content": "Say hello"}] + assert response.choices[0].message.content == "Hello from VoltageGPU"