From f9dec7e45d3907b21c8e2b81698b5ace6cd42e22 Mon Sep 17 00:00:00 2001 From: Robert Keus Date: Sun, 26 Jul 2026 10:29:47 +0200 Subject: [PATCH 01/10] feat(provider): add GreenPT support --- litellm/llms/openai_like/providers.json | 9 ++ litellm/types/utils.py | 1 + model_prices_and_context_window.json | 123 ++++++++++++++++++ provider_endpoints_support.json | 17 +++ .../llms/openai_like/test_greenpt_provider.py | 71 ++++++++++ 5 files changed, 221 insertions(+) create mode 100644 tests/litellm/llms/openai_like/test_greenpt_provider.py diff --git a/litellm/llms/openai_like/providers.json b/litellm/llms/openai_like/providers.json index 164100d4194..acb5f7d12bd 100644 --- a/litellm/llms/openai_like/providers.json +++ b/litellm/llms/openai_like/providers.json @@ -168,6 +168,15 @@ }, "supported_endpoints": ["/v1/chat/completions", "/v1/responses"] }, + "greenpt": { + "base_url": "https://api.greenpt.ai/v1", + "api_key_env": "GREENPT_API_KEY", + "api_base_env": "GREENPT_API_BASE", + "param_mappings": { + "max_completion_tokens": "max_tokens" + }, + "supported_endpoints": ["/v1/chat/completions", "/v1/embeddings"] + }, "meta": { "base_url": "https://api.meta.ai/v1", "api_key_env": "META_API_KEY", diff --git a/litellm/types/utils.py b/litellm/types/utils.py index e4dfac48141..2470c6b600b 100644 --- a/litellm/types/utils.py +++ b/litellm/types/utils.py @@ -3467,6 +3467,7 @@ class LlmProviders(str, Enum): TOPAZ = "topaz" SAP_GENERATIVE_AI_HUB = "sap" ASSEMBLYAI = "assemblyai" + GREENPT = "greenpt" CHARITY_ENGINE = "charity_engine" GITHUB_COPILOT = "github_copilot" SNOWFLAKE = "snowflake" diff --git a/model_prices_and_context_window.json b/model_prices_and_context_window.json index db28118d52b..2ac7b06c43d 100644 --- a/model_prices_and_context_window.json +++ b/model_prices_and_context_window.json @@ -46260,6 +46260,129 @@ "supports_system_messages": true, "supports_tool_choice": true }, + "greenpt/gemma4": { + "input_cost_per_token": 5.6885e-07, + "litellm_provider": "greenpt", + "max_input_tokens": 262144, + "max_output_tokens": 32768, + "max_tokens": 32768, + "mode": "chat", + "output_cost_per_token": 1.70655e-06, + "source": "https://docs.greenpt.ai/pricing", + "supports_reasoning": true, + "supports_vision": true + }, + "greenpt/glm-5.2": { + "input_cost_per_token": 1.501764e-06, + "litellm_provider": "greenpt", + "max_input_tokens": 1000000, + "max_output_tokens": 131072, + "max_tokens": 131072, + "mode": "chat", + "output_cost_per_token": 5.256174e-06, + "source": "https://docs.greenpt.ai/pricing", + "supports_function_calling": true, + "supports_reasoning": true, + "supports_tool_choice": true + }, + "greenpt/green-embedding": { + "input_cost_per_token": 2.2754e-07, + "litellm_provider": "greenpt", + "max_input_tokens": 32768, + "max_output_tokens": 2560, + "max_tokens": 32768, + "mode": "embedding", + "output_cost_per_token": 0, + "source": "https://docs.greenpt.ai/pricing" + }, + "greenpt/green-l": { + "input_cost_per_token": 2.84425e-07, + "litellm_provider": "greenpt", + "max_input_tokens": 128000, + "max_output_tokens": 32000, + "max_tokens": 32000, + "mode": "chat", + "output_cost_per_token": 9.1016e-07, + "source": "https://docs.greenpt.ai/pricing", + "supports_function_calling": true, + "supports_tool_choice": true, + "supports_vision": true + }, + "greenpt/green-l-raw": { + "input_cost_per_token": 2.84425e-07, + "litellm_provider": "greenpt", + "max_input_tokens": 128000, + "max_output_tokens": 32000, + "max_tokens": 32000, + "mode": "chat", + "output_cost_per_token": 9.1016e-07, + "source": "https://docs.greenpt.ai/pricing", + "supports_function_calling": true, + "supports_tool_choice": true, + "supports_vision": true + }, + "greenpt/green-r": { + "input_cost_per_token": 3.98195e-07, + "litellm_provider": "greenpt", + "max_input_tokens": 131072, + "max_output_tokens": 32768, + "max_tokens": 32768, + "mode": "chat", + "output_cost_per_token": 1.080815e-06, + "source": "https://docs.greenpt.ai/pricing", + "supports_function_calling": true, + "supports_reasoning": true, + "supports_tool_choice": true, + "supports_vision": true + }, + "greenpt/green-r-raw": { + "input_cost_per_token": 3.98195e-07, + "litellm_provider": "greenpt", + "max_input_tokens": 131072, + "max_output_tokens": 32768, + "max_tokens": 32768, + "mode": "chat", + "output_cost_per_token": 1.080815e-06, + "source": "https://docs.greenpt.ai/pricing", + "supports_function_calling": true, + "supports_reasoning": true, + "supports_tool_choice": true, + "supports_vision": true + }, + "greenpt/kimi-k2.6": { + "input_cost_per_token": 8.30521e-07, + "litellm_provider": "greenpt", + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "max_tokens": 262144, + "mode": "chat", + "output_cost_per_token": 4.380145e-06, + "source": "https://docs.greenpt.ai/pricing", + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_reasoning": true, + "supports_tool_choice": true, + "supports_vision": true + }, + "greenpt/minimax-m2.5": { + "input_cost_per_token": 1.93409e-07, + "litellm_provider": "greenpt", + "max_input_tokens": 200000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 1.126323e-06, + "source": "https://docs.greenpt.ai/pricing", + "supports_function_calling": true, + "supports_reasoning": true, + "supports_tool_choice": true + }, "fallback_generalizations": { "rules": [ { diff --git a/provider_endpoints_support.json b/provider_endpoints_support.json index 65db63dc045..ec1baed2292 100644 --- a/provider_endpoints_support.json +++ b/provider_endpoints_support.json @@ -2941,6 +2941,23 @@ "rerank": false, "a2a": false } + }, + "greenpt": { + "display_name": "GreenPT (`greenpt`)", + "url": "https://docs.greenpt.ai", + "endpoints": { + "chat_completions": true, + "messages": false, + "responses": false, + "embeddings": true, + "image_generations": false, + "audio_transcriptions": false, + "audio_speech": false, + "moderations": false, + "batches": false, + "rerank": false, + "a2a": false + } } }, "endpoints": { diff --git a/tests/litellm/llms/openai_like/test_greenpt_provider.py b/tests/litellm/llms/openai_like/test_greenpt_provider.py new file mode 100644 index 00000000000..648f6f2d90b --- /dev/null +++ b/tests/litellm/llms/openai_like/test_greenpt_provider.py @@ -0,0 +1,71 @@ +import os +import sys + +sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), "../../../.."))) + +from litellm.llms.openai_like.dynamic_config import create_config_class +from litellm.llms.openai_like.json_loader import JSONProviderRegistry + +GREENPT_BASE_URL = "https://api.greenpt.ai/v1" + + +def _get_config(): + provider = JSONProviderRegistry.get("greenpt") + assert provider is not None + config_class = create_config_class(provider) + return config_class() + + +def test_greenpt_provider_registered(): + provider = JSONProviderRegistry.get("greenpt") + assert provider is not None + assert provider.base_url == GREENPT_BASE_URL + assert provider.api_key_env == "GREENPT_API_KEY" + assert provider.api_base_env == "GREENPT_API_BASE" + + +def test_greenpt_resolves_env_api_key(monkeypatch): + config = _get_config() + monkeypatch.setenv("GREENPT_API_KEY", "test-key") + api_base, api_key = config._get_openai_compatible_provider_info(None, None) + assert api_base == GREENPT_BASE_URL + assert api_key == "test-key" + + +def test_greenpt_maps_max_completion_tokens(): + config = _get_config() + params = config.map_openai_params( + non_default_params={"max_completion_tokens": 256}, + optional_params={}, + model="greenpt/gemma4", + drop_params=False, + ) + assert params == {"max_tokens": 256} + + +def test_greenpt_complete_url_appends_endpoint(): + config = _get_config() + url = config.get_complete_url( + api_base=GREENPT_BASE_URL, + api_key="test-key", + model="greenpt/gemma4", + optional_params={}, + litellm_params={}, + stream=False, + ) + assert url == f"{GREENPT_BASE_URL}/chat/completions" + + +def test_greenpt_provider_resolution(): + from litellm.litellm_core_utils.get_llm_provider_logic import get_llm_provider + + model, provider, _, api_base = get_llm_provider( + model="greenpt/gemma4", + custom_llm_provider=None, + api_base=None, + api_key=None, + ) + + assert model == "gemma4" + assert provider == "greenpt" + assert api_base == GREENPT_BASE_URL From 2d2124cfce3104d6b036e1c1f277bd6cc2802ea8 Mon Sep 17 00:00:00 2001 From: Robert Keus Date: Sun, 26 Jul 2026 11:47:46 +0200 Subject: [PATCH 02/10] feat(greenpt): add flagship and rerank support --- litellm/llms/greenpt/__init__.py | 1 + litellm/llms/greenpt/rerank/__init__.py | 1 + litellm/llms/greenpt/rerank/transformation.py | 42 +++++++++++++++++++ litellm/rerank_api/main.py | 3 +- litellm/utils.py | 4 ++ model_prices_and_context_window.json | 31 ++++++++++++++ provider_endpoints_support.json | 2 +- .../llms/openai_like/test_greenpt_provider.py | 27 ++++++++++++ 8 files changed, 109 insertions(+), 2 deletions(-) create mode 100644 litellm/llms/greenpt/__init__.py create mode 100644 litellm/llms/greenpt/rerank/__init__.py create mode 100644 litellm/llms/greenpt/rerank/transformation.py diff --git a/litellm/llms/greenpt/__init__.py b/litellm/llms/greenpt/__init__.py new file mode 100644 index 00000000000..8b137891791 --- /dev/null +++ b/litellm/llms/greenpt/__init__.py @@ -0,0 +1 @@ + diff --git a/litellm/llms/greenpt/rerank/__init__.py b/litellm/llms/greenpt/rerank/__init__.py new file mode 100644 index 00000000000..8b137891791 --- /dev/null +++ b/litellm/llms/greenpt/rerank/__init__.py @@ -0,0 +1 @@ + diff --git a/litellm/llms/greenpt/rerank/transformation.py b/litellm/llms/greenpt/rerank/transformation.py new file mode 100644 index 00000000000..8e2744ba537 --- /dev/null +++ b/litellm/llms/greenpt/rerank/transformation.py @@ -0,0 +1,42 @@ +from typing import Dict + +from litellm.llms.hosted_vllm.rerank.transformation import HostedVLLMRerankConfig +from litellm.secret_managers.main import get_secret_str + + +class GreenPTRerankConfig(HostedVLLMRerankConfig): + def get_supported_cohere_rerank_params(self, model: str) -> list: + return ["query", "documents", "top_n", "return_documents"] + + def validate_environment( + self, + headers: dict, + model: str, + api_key: str | None = None, + optional_params: dict | None = None, + ) -> dict: + api_key = api_key or get_secret_str("GREENPT_API_KEY") + if api_key is None: + raise ValueError("GreenPT API key is required. Set GREENPT_API_KEY.") + return super().validate_environment( + headers=headers, + model=model, + api_key=api_key, + optional_params=optional_params, + ) + + def transform_rerank_request( + self, + model: str, + optional_rerank_params: Dict, + headers: dict, + litellm_params: dict | None = None, + ) -> dict: + request = super().transform_rerank_request( + model=model, + optional_rerank_params=optional_rerank_params, + headers=headers, + litellm_params=litellm_params, + ) + request.setdefault("top_n", len(request["documents"])) + return request diff --git a/litellm/rerank_api/main.py b/litellm/rerank_api/main.py index b6ebf5589b7..63c0df3c48e 100644 --- a/litellm/rerank_api/main.py +++ b/litellm/rerank_api/main.py @@ -31,7 +31,7 @@ async def arerank( query: str, documents: List[Union[str, Dict[str, Any]]], custom_llm_provider: ( - Literal["cohere", "together_ai", "deepinfra", "fireworks_ai", "voyage", "watsonx"] | None + Literal["cohere", "together_ai", "deepinfra", "fireworks_ai", "voyage", "watsonx", "greenpt"] | None ) = None, top_n: int | None = None, rank_fields: List[str] | None = None, @@ -89,6 +89,7 @@ def rerank( "fireworks_ai", "voyage", "watsonx", + "greenpt", ] | None ) = None, diff --git a/litellm/utils.py b/litellm/utils.py index a11c5500503..7776eec6698 100644 --- a/litellm/utils.py +++ b/litellm/utils.py @@ -8008,6 +8008,10 @@ class ProviderConfigManager: ) return DashScopeRerankConfig() + elif litellm.LlmProviders.GREENPT == provider: + from litellm.llms.greenpt.rerank.transformation import GreenPTRerankConfig + + return GreenPTRerankConfig() return litellm.CohereRerankConfig() @staticmethod diff --git a/model_prices_and_context_window.json b/model_prices_and_context_window.json index 2ac7b06c43d..00061889e35 100644 --- a/model_prices_and_context_window.json +++ b/model_prices_and_context_window.json @@ -46295,6 +46295,14 @@ "output_cost_per_token": 0, "source": "https://docs.greenpt.ai/pricing" }, + "greenpt/green-rerank": { + "input_cost_per_token": 1.36524e-07, + "litellm_provider": "greenpt", + "max_input_tokens": 32768, + "max_tokens": 32768, + "mode": "rerank", + "source": "https://docs.greenpt.ai/pricing" + }, "greenpt/green-l": { "input_cost_per_token": 2.84425e-07, "litellm_provider": "greenpt", @@ -46370,6 +46378,19 @@ "supports_tool_choice": true, "supports_vision": true }, + "greenpt/kimi-k2.7-code": { + "input_cost_per_token": 9.44291e-07, + "litellm_provider": "greenpt", + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "max_tokens": 262144, + "mode": "chat", + "output_cost_per_token": 4.380145e-06, + "source": "https://docs.greenpt.ai/pricing", + "supports_function_calling": true, + "supports_reasoning": true, + "supports_tool_choice": true + }, "greenpt/minimax-m2.5": { "input_cost_per_token": 1.93409e-07, "litellm_provider": "greenpt", @@ -46383,6 +46404,16 @@ "supports_reasoning": true, "supports_tool_choice": true }, + "greenpt/qwen3-embedding-8b": { + "input_cost_per_token": 2.84425e-07, + "litellm_provider": "greenpt", + "max_input_tokens": 32768, + "max_output_tokens": 4096, + "max_tokens": 32768, + "mode": "embedding", + "output_cost_per_token": 0, + "source": "https://docs.greenpt.ai/pricing" + }, "fallback_generalizations": { "rules": [ { diff --git a/provider_endpoints_support.json b/provider_endpoints_support.json index ec1baed2292..ce52bbde3f8 100644 --- a/provider_endpoints_support.json +++ b/provider_endpoints_support.json @@ -2955,7 +2955,7 @@ "audio_speech": false, "moderations": false, "batches": false, - "rerank": false, + "rerank": true, "a2a": false } } diff --git a/tests/litellm/llms/openai_like/test_greenpt_provider.py b/tests/litellm/llms/openai_like/test_greenpt_provider.py index 648f6f2d90b..21a794657e7 100644 --- a/tests/litellm/llms/openai_like/test_greenpt_provider.py +++ b/tests/litellm/llms/openai_like/test_greenpt_provider.py @@ -5,6 +5,8 @@ sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), "../. from litellm.llms.openai_like.dynamic_config import create_config_class from litellm.llms.openai_like.json_loader import JSONProviderRegistry +from litellm.types.utils import LlmProviders +from litellm.utils import ProviderConfigManager GREENPT_BASE_URL = "https://api.greenpt.ai/v1" @@ -69,3 +71,28 @@ def test_greenpt_provider_resolution(): assert model == "gemma4" assert provider == "greenpt" assert api_base == GREENPT_BASE_URL + + +def test_greenpt_rerank_config(): + config = ProviderConfigManager.get_provider_rerank_config( + model="green-rerank", + provider=LlmProviders.GREENPT, + api_base=GREENPT_BASE_URL, + present_version_params=[], + ) + assert config.get_complete_url(GREENPT_BASE_URL, "green-rerank") == (f"{GREENPT_BASE_URL}/rerank") + assert config.transform_rerank_request( + model="green-rerank", + optional_rerank_params={ + "query": "low-carbon inference", + "documents": ["GreenPT", "Other provider"], + "return_documents": False, + }, + headers={}, + ) == { + "model": "green-rerank", + "query": "low-carbon inference", + "documents": ["GreenPT", "Other provider"], + "top_n": 2, + "return_documents": False, + } From 015a2f24b4813f206680f2bb0f92cf6a4c6f5379 Mon Sep 17 00:00:00 2001 From: Robert Keus Date: Sun, 26 Jul 2026 16:14:04 +0200 Subject: [PATCH 03/10] fix(greenpt): pass provider CI checks --- litellm/llms/greenpt/rerank/transformation.py | 4 +- litellm/utils.py | 9 +-- model_prices_and_context_window.json | 4 +- .../greenpt/rerank/test_transformation.py | 66 +++++++++++++++++++ .../llms/openai_like/test_greenpt_provider.py | 15 ----- 5 files changed, 74 insertions(+), 24 deletions(-) create mode 100644 tests/litellm/llms/greenpt/rerank/test_transformation.py diff --git a/litellm/llms/greenpt/rerank/transformation.py b/litellm/llms/greenpt/rerank/transformation.py index 8e2744ba537..54daa1fe4c7 100644 --- a/litellm/llms/greenpt/rerank/transformation.py +++ b/litellm/llms/greenpt/rerank/transformation.py @@ -1,5 +1,3 @@ -from typing import Dict - from litellm.llms.hosted_vllm.rerank.transformation import HostedVLLMRerankConfig from litellm.secret_managers.main import get_secret_str @@ -28,7 +26,7 @@ class GreenPTRerankConfig(HostedVLLMRerankConfig): def transform_rerank_request( self, model: str, - optional_rerank_params: Dict, + optional_rerank_params: dict, headers: dict, litellm_params: dict | None = None, ) -> dict: diff --git a/litellm/utils.py b/litellm/utils.py index 7776eec6698..a16d5a093aa 100644 --- a/litellm/utils.py +++ b/litellm/utils.py @@ -8008,11 +8008,12 @@ class ProviderConfigManager: ) return DashScopeRerankConfig() - elif litellm.LlmProviders.GREENPT == provider: - from litellm.llms.greenpt.rerank.transformation import GreenPTRerankConfig + from litellm.llms.greenpt.rerank.transformation import GreenPTRerankConfig - return GreenPTRerankConfig() - return litellm.CohereRerankConfig() + config_class = { + litellm.LlmProviders.GREENPT: GreenPTRerankConfig, + }.get(provider, litellm.CohereRerankConfig) + return config_class() @staticmethod def get_provider_anthropic_messages_config( diff --git a/model_prices_and_context_window.json b/model_prices_and_context_window.json index 00061889e35..20940e843a7 100644 --- a/model_prices_and_context_window.json +++ b/model_prices_and_context_window.json @@ -46290,7 +46290,7 @@ "litellm_provider": "greenpt", "max_input_tokens": 32768, "max_output_tokens": 2560, - "max_tokens": 32768, + "max_tokens": 2560, "mode": "embedding", "output_cost_per_token": 0, "source": "https://docs.greenpt.ai/pricing" @@ -46409,7 +46409,7 @@ "litellm_provider": "greenpt", "max_input_tokens": 32768, "max_output_tokens": 4096, - "max_tokens": 32768, + "max_tokens": 4096, "mode": "embedding", "output_cost_per_token": 0, "source": "https://docs.greenpt.ai/pricing" diff --git a/tests/litellm/llms/greenpt/rerank/test_transformation.py b/tests/litellm/llms/greenpt/rerank/test_transformation.py new file mode 100644 index 00000000000..cf12bfd80fd --- /dev/null +++ b/tests/litellm/llms/greenpt/rerank/test_transformation.py @@ -0,0 +1,66 @@ +import pytest + +from litellm.llms.greenpt.rerank.transformation import GreenPTRerankConfig + + +def test_supported_params(): + assert GreenPTRerankConfig().get_supported_cohere_rerank_params("green-rerank") == [ + "query", + "documents", + "top_n", + "return_documents", + ] + + +def test_validate_environment_with_api_key(): + headers = GreenPTRerankConfig().validate_environment( + headers={}, + model="green-rerank", + api_key="test-key", + ) + assert headers == { + "Authorization": "Bearer test-key", + "accept": "application/json", + "content-type": "application/json", + } + + +def test_validate_environment_requires_api_key(monkeypatch): + monkeypatch.delenv("GREENPT_API_KEY", raising=False) + with pytest.raises(ValueError, match="GreenPT API key is required"): + GreenPTRerankConfig().validate_environment( + headers={}, + model="green-rerank", + ) + + +def test_transform_rerank_request_defaults_top_n(): + request = GreenPTRerankConfig().transform_rerank_request( + model="green-rerank", + optional_rerank_params={ + "query": "low-carbon inference", + "documents": ["GreenPT", "Other provider"], + "return_documents": False, + }, + headers={}, + ) + assert request == { + "model": "green-rerank", + "query": "low-carbon inference", + "documents": ["GreenPT", "Other provider"], + "top_n": 2, + "return_documents": False, + } + + +def test_transform_rerank_request_preserves_top_n(): + request = GreenPTRerankConfig().transform_rerank_request( + model="green-rerank", + optional_rerank_params={ + "query": "low-carbon inference", + "documents": ["GreenPT", "Other provider"], + "top_n": 1, + }, + headers={}, + ) + assert request["top_n"] == 1 diff --git a/tests/litellm/llms/openai_like/test_greenpt_provider.py b/tests/litellm/llms/openai_like/test_greenpt_provider.py index 21a794657e7..efe5f2808b9 100644 --- a/tests/litellm/llms/openai_like/test_greenpt_provider.py +++ b/tests/litellm/llms/openai_like/test_greenpt_provider.py @@ -81,18 +81,3 @@ def test_greenpt_rerank_config(): present_version_params=[], ) assert config.get_complete_url(GREENPT_BASE_URL, "green-rerank") == (f"{GREENPT_BASE_URL}/rerank") - assert config.transform_rerank_request( - model="green-rerank", - optional_rerank_params={ - "query": "low-carbon inference", - "documents": ["GreenPT", "Other provider"], - "return_documents": False, - }, - headers={}, - ) == { - "model": "green-rerank", - "query": "low-carbon inference", - "documents": ["GreenPT", "Other provider"], - "top_n": 2, - "return_documents": False, - } From 5288ceaca5131d94ec4cb817dc95de99acf39353 Mon Sep 17 00:00:00 2001 From: Robert Keus Date: Mon, 27 Jul 2026 08:44:47 +0200 Subject: [PATCH 04/10] fix(greenpt): enable embeddings and rerank cost tracking --- litellm/llms/greenpt/rerank/transformation.py | 19 +++++++++ litellm/llms/openai_like/json_loader.py | 7 ++++ litellm/main.py | 1 + .../greenpt/rerank/test_transformation.py | 26 +++++++++++++ .../llms/openai_like/test_greenpt_provider.py | 39 +++++++++++++++++-- 5 files changed, 88 insertions(+), 4 deletions(-) rename tests/{litellm => test_litellm}/llms/greenpt/rerank/test_transformation.py (70%) rename tests/{litellm => test_litellm}/llms/openai_like/test_greenpt_provider.py (67%) diff --git a/litellm/llms/greenpt/rerank/transformation.py b/litellm/llms/greenpt/rerank/transformation.py index 54daa1fe4c7..b36c3029dae 100644 --- a/litellm/llms/greenpt/rerank/transformation.py +++ b/litellm/llms/greenpt/rerank/transformation.py @@ -1,5 +1,7 @@ from litellm.llms.hosted_vllm.rerank.transformation import HostedVLLMRerankConfig from litellm.secret_managers.main import get_secret_str +from litellm.types.rerank import RerankBilledUnits +from litellm.types.utils import ModelInfo class GreenPTRerankConfig(HostedVLLMRerankConfig): @@ -38,3 +40,20 @@ class GreenPTRerankConfig(HostedVLLMRerankConfig): ) request.setdefault("top_n", len(request["documents"])) return request + + def calculate_rerank_cost( + self, + model: str, + custom_llm_provider: str | None = None, + billed_units: RerankBilledUnits | None = None, + model_info: ModelInfo | None = None, + ) -> tuple[float, float]: + if model_info is None or billed_units is None: + return 0.0, 0.0 + input_cost_per_token = model_info.get("input_cost_per_token") + if input_cost_per_token is None: + return 0.0, 0.0 + total_tokens = billed_units.get("total_tokens") + if total_tokens is None: + return 0.0, 0.0 + return input_cost_per_token * total_tokens, 0.0 diff --git a/litellm/llms/openai_like/json_loader.py b/litellm/llms/openai_like/json_loader.py index 4640bb8a422..451e05364a1 100644 --- a/litellm/llms/openai_like/json_loader.py +++ b/litellm/llms/openai_like/json_loader.py @@ -73,6 +73,13 @@ class JSONProviderRegistry: return False return "/v1/responses" in provider.supported_endpoints + @classmethod + def supports_embeddings_api(cls, slug: str) -> bool: + provider = cls._providers.get(slug) + if provider is None: + return False + return "/v1/embeddings" in provider.supported_endpoints + @classmethod def list_providers(cls) -> list: """List all registered provider slugs""" diff --git a/litellm/main.py b/litellm/main.py index dc3ec469a1b..c08efdf76f9 100644 --- a/litellm/main.py +++ b/litellm/main.py @@ -6213,6 +6213,7 @@ def embedding( custom_llm_provider == "openai_like" or custom_llm_provider == "llamafile" or custom_llm_provider == "lm_studio" + or JSONProviderRegistry.supports_embeddings_api(custom_llm_provider) ): api_base = api_base or litellm.api_base or get_secret_str("OPENAI_LIKE_API_BASE") diff --git a/tests/litellm/llms/greenpt/rerank/test_transformation.py b/tests/test_litellm/llms/greenpt/rerank/test_transformation.py similarity index 70% rename from tests/litellm/llms/greenpt/rerank/test_transformation.py rename to tests/test_litellm/llms/greenpt/rerank/test_transformation.py index cf12bfd80fd..9ae9df8edcf 100644 --- a/tests/litellm/llms/greenpt/rerank/test_transformation.py +++ b/tests/test_litellm/llms/greenpt/rerank/test_transformation.py @@ -64,3 +64,29 @@ def test_transform_rerank_request_preserves_top_n(): headers={}, ) assert request["top_n"] == 1 + + +def test_calculate_rerank_cost(): + prompt_cost, completion_cost = GreenPTRerankConfig().calculate_rerank_cost( + model="green-rerank", + billed_units={"total_tokens": 1000}, + model_info={"input_cost_per_token": 1.36524e-07}, + ) + assert prompt_cost == pytest.approx(0.000136524) + assert completion_cost == 0.0 + + +@pytest.mark.parametrize( + ("billed_units", "model_info"), + [ + (None, None), + ({}, {"input_cost_per_token": 1.36524e-07}), + ({"total_tokens": 1000}, {}), + ], +) +def test_calculate_rerank_cost_missing_usage_or_pricing(billed_units, model_info): + assert GreenPTRerankConfig().calculate_rerank_cost( + model="green-rerank", + billed_units=billed_units, + model_info=model_info, + ) == (0.0, 0.0) diff --git a/tests/litellm/llms/openai_like/test_greenpt_provider.py b/tests/test_litellm/llms/openai_like/test_greenpt_provider.py similarity index 67% rename from tests/litellm/llms/openai_like/test_greenpt_provider.py rename to tests/test_litellm/llms/openai_like/test_greenpt_provider.py index efe5f2808b9..bf78b5e94be 100644 --- a/tests/litellm/llms/openai_like/test_greenpt_provider.py +++ b/tests/test_litellm/llms/openai_like/test_greenpt_provider.py @@ -1,8 +1,12 @@ +import json import os import sys +from unittest.mock import Mock, patch sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), "../../../.."))) +import litellm +from litellm.llms.custom_httpx.http_handler import HTTPHandler from litellm.llms.openai_like.dynamic_config import create_config_class from litellm.llms.openai_like.json_loader import JSONProviderRegistry from litellm.types.utils import LlmProviders @@ -39,7 +43,7 @@ def test_greenpt_maps_max_completion_tokens(): params = config.map_openai_params( non_default_params={"max_completion_tokens": 256}, optional_params={}, - model="greenpt/gemma4", + model="greenpt/glm-5.2", drop_params=False, ) assert params == {"max_tokens": 256} @@ -50,7 +54,7 @@ def test_greenpt_complete_url_appends_endpoint(): url = config.get_complete_url( api_base=GREENPT_BASE_URL, api_key="test-key", - model="greenpt/gemma4", + model="greenpt/glm-5.2", optional_params={}, litellm_params={}, stream=False, @@ -62,17 +66,44 @@ def test_greenpt_provider_resolution(): from litellm.litellm_core_utils.get_llm_provider_logic import get_llm_provider model, provider, _, api_base = get_llm_provider( - model="greenpt/gemma4", + model="greenpt/glm-5.2", custom_llm_provider=None, api_base=None, api_key=None, ) - assert model == "gemma4" + assert model == "glm-5.2" assert provider == "greenpt" assert api_base == GREENPT_BASE_URL +def test_greenpt_embedding_dispatches(): + response = Mock() + response.raise_for_status = Mock() + response.json.return_value = { + "object": "list", + "data": [{"object": "embedding", "embedding": [0.1, 0.2], "index": 0}], + "model": "green-embedding", + "usage": {"prompt_tokens": 2, "total_tokens": 2}, + } + client = HTTPHandler() + + with patch.object(HTTPHandler, "post", return_value=response) as mock_post: + result = litellm.embedding( + model="greenpt/green-embedding", + input=["renewable inference"], + api_key="test-key", + client=client, + ) + + assert result.data[0]["embedding"] == [0.1, 0.2] + assert mock_post.call_args.args[0] == f"{GREENPT_BASE_URL}/embeddings" + assert json.loads(mock_post.call_args.kwargs["data"]) == { + "model": "green-embedding", + "input": ["renewable inference"], + } + + def test_greenpt_rerank_config(): config = ProviderConfigManager.get_provider_rerank_config( model="green-rerank", From 49e568a978bb756665251eba17019abc1789baf7 Mon Sep 17 00:00:00 2001 From: Robert Keus Date: Mon, 27 Jul 2026 08:56:50 +0200 Subject: [PATCH 05/10] test(greenpt): avoid provider test module collision --- ..._transformation.py => test_greenpt_rerank_transformation.py} | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) rename tests/test_litellm/llms/greenpt/rerank/{test_transformation.py => test_greenpt_rerank_transformation.py} (98%) diff --git a/tests/test_litellm/llms/greenpt/rerank/test_transformation.py b/tests/test_litellm/llms/greenpt/rerank/test_greenpt_rerank_transformation.py similarity index 98% rename from tests/test_litellm/llms/greenpt/rerank/test_transformation.py rename to tests/test_litellm/llms/greenpt/rerank/test_greenpt_rerank_transformation.py index 9ae9df8edcf..0902acd09e8 100644 --- a/tests/test_litellm/llms/greenpt/rerank/test_transformation.py +++ b/tests/test_litellm/llms/greenpt/rerank/test_greenpt_rerank_transformation.py @@ -3,7 +3,7 @@ import pytest from litellm.llms.greenpt.rerank.transformation import GreenPTRerankConfig -def test_supported_params(): +def test_greenpt_supported_params(): assert GreenPTRerankConfig().get_supported_cohere_rerank_params("green-rerank") == [ "query", "documents", From 7d1ea76130c4bbdd7f8f7521b2ef336e427bd26f Mon Sep 17 00:00:00 2001 From: Robert Keus Date: Mon, 27 Jul 2026 19:12:10 +0200 Subject: [PATCH 06/10] test(greenpt): cover public rerank cost path --- .../test_greenpt_rerank_transformation.py | 28 +++++++++++++++++++ 1 file changed, 28 insertions(+) diff --git a/tests/test_litellm/llms/greenpt/rerank/test_greenpt_rerank_transformation.py b/tests/test_litellm/llms/greenpt/rerank/test_greenpt_rerank_transformation.py index 0902acd09e8..5eb6c7c0c7a 100644 --- a/tests/test_litellm/llms/greenpt/rerank/test_greenpt_rerank_transformation.py +++ b/tests/test_litellm/llms/greenpt/rerank/test_greenpt_rerank_transformation.py @@ -1,8 +1,26 @@ +import json +from pathlib import Path + import pytest +import litellm +from litellm.cost_calculator import rerank_cost from litellm.llms.greenpt.rerank.transformation import GreenPTRerankConfig +@pytest.fixture +def local_model_cost_map(): + original_model_cost = litellm.model_cost + model_cost_path = Path(__file__).resolve().parents[5] / "model_prices_and_context_window.json" + litellm.model_cost = json.loads(model_cost_path.read_text(encoding="utf-8")) + litellm.get_model_info.cache_clear() + try: + yield + finally: + litellm.model_cost = original_model_cost + litellm.get_model_info.cache_clear() + + def test_greenpt_supported_params(): assert GreenPTRerankConfig().get_supported_cohere_rerank_params("green-rerank") == [ "query", @@ -76,6 +94,16 @@ def test_calculate_rerank_cost(): assert completion_cost == 0.0 +def test_public_rerank_cost_uses_token_pricing(local_model_cost_map): + prompt_cost, completion_cost = rerank_cost( + model="greenpt/green-rerank", + custom_llm_provider=None, + billed_units={"total_tokens": 1000}, + ) + assert prompt_cost == pytest.approx(0.000136524) + assert completion_cost == 0.0 + + @pytest.mark.parametrize( ("billed_units", "model_info"), [ From 5c55c88dcd0739f3fa61830ee057002d9e0b6ba2 Mon Sep 17 00:00:00 2001 From: Robert Keus Date: Tue, 28 Jul 2026 06:46:49 +0200 Subject: [PATCH 07/10] chore(greenpt): focus model catalog on flagship models --- model_prices_and_context_window.json | 12 ------------ 1 file changed, 12 deletions(-) diff --git a/model_prices_and_context_window.json b/model_prices_and_context_window.json index 20940e843a7..7ee81303c47 100644 --- a/model_prices_and_context_window.json +++ b/model_prices_and_context_window.json @@ -46260,18 +46260,6 @@ "supports_system_messages": true, "supports_tool_choice": true }, - "greenpt/gemma4": { - "input_cost_per_token": 5.6885e-07, - "litellm_provider": "greenpt", - "max_input_tokens": 262144, - "max_output_tokens": 32768, - "max_tokens": 32768, - "mode": "chat", - "output_cost_per_token": 1.70655e-06, - "source": "https://docs.greenpt.ai/pricing", - "supports_reasoning": true, - "supports_vision": true - }, "greenpt/glm-5.2": { "input_cost_per_token": 1.501764e-06, "litellm_provider": "greenpt", From c31b7da4ffb1d879a98965a9d6aa46aaba0b9ef8 Mon Sep 17 00:00:00 2001 From: Robert Keus Date: Sun, 11 Oct 2026 08:27:42 +0200 Subject: [PATCH 08/10] fix(greenpt): satisfy type-discipline and test-quality gates Co-Authored-By: Claude Opus 5.5 --- litellm/llms/greenpt/rerank/transformation.py | 30 ++++++++++--------- litellm/llms/openai_like/json_loader.py | 2 +- litellm/utils.py | 2 +- .../test_greenpt_rerank_transformation.py | 18 +---------- .../llms/openai_like/test_greenpt_provider.py | 4 --- 5 files changed, 19 insertions(+), 37 deletions(-) diff --git a/litellm/llms/greenpt/rerank/transformation.py b/litellm/llms/greenpt/rerank/transformation.py index b36c3029dae..d62c5d5538f 100644 --- a/litellm/llms/greenpt/rerank/transformation.py +++ b/litellm/llms/greenpt/rerank/transformation.py @@ -1,3 +1,5 @@ +from typing import Final + from litellm.llms.hosted_vllm.rerank.transformation import HostedVLLMRerankConfig from litellm.secret_managers.main import get_secret_str from litellm.types.rerank import RerankBilledUnits @@ -5,34 +7,34 @@ from litellm.types.utils import ModelInfo class GreenPTRerankConfig(HostedVLLMRerankConfig): - def get_supported_cohere_rerank_params(self, model: str) -> list: + def get_supported_cohere_rerank_params(self, model: str) -> list: # mutable-ok: mirrors override contract return ["query", "documents", "top_n", "return_documents"] def validate_environment( self, - headers: dict, + headers: dict, # mutable-ok: mirrors override contract model: str, api_key: str | None = None, - optional_params: dict | None = None, - ) -> dict: - api_key = api_key or get_secret_str("GREENPT_API_KEY") - if api_key is None: + optional_params: dict | None = None, # mutable-ok: mirrors override contract + ) -> dict: # mutable-ok: mirrors override contract + resolved_api_key: Final = api_key or get_secret_str("GREENPT_API_KEY") + if resolved_api_key is None: raise ValueError("GreenPT API key is required. Set GREENPT_API_KEY.") return super().validate_environment( headers=headers, model=model, - api_key=api_key, + api_key=resolved_api_key, optional_params=optional_params, ) def transform_rerank_request( self, model: str, - optional_rerank_params: dict, - headers: dict, - litellm_params: dict | None = None, - ) -> dict: - request = super().transform_rerank_request( + optional_rerank_params: dict, # mutable-ok: mirrors override contract + headers: dict, # mutable-ok: mirrors override contract + litellm_params: dict | None = None, # mutable-ok: mirrors override contract + ) -> dict: # mutable-ok: mirrors override contract + request: Final = super().transform_rerank_request( model=model, optional_rerank_params=optional_rerank_params, headers=headers, @@ -50,10 +52,10 @@ class GreenPTRerankConfig(HostedVLLMRerankConfig): ) -> tuple[float, float]: if model_info is None or billed_units is None: return 0.0, 0.0 - input_cost_per_token = model_info.get("input_cost_per_token") + input_cost_per_token: Final = model_info.get("input_cost_per_token") if input_cost_per_token is None: return 0.0, 0.0 - total_tokens = billed_units.get("total_tokens") + total_tokens: Final = billed_units.get("total_tokens") if total_tokens is None: return 0.0, 0.0 return input_cost_per_token * total_tokens, 0.0 diff --git a/litellm/llms/openai_like/json_loader.py b/litellm/llms/openai_like/json_loader.py index c0addb68db3..7bbe523b0c0 100644 --- a/litellm/llms/openai_like/json_loader.py +++ b/litellm/llms/openai_like/json_loader.py @@ -80,7 +80,7 @@ class JSONProviderRegistry: @classmethod def supports_embeddings_api(cls, slug: str) -> bool: - provider = cls._providers.get(slug) + provider: Final = cls._providers.get(slug) if provider is None: return False return "/v1/embeddings" in provider.supported_endpoints diff --git a/litellm/utils.py b/litellm/utils.py index 6724229d24b..61af22bb5c0 100644 --- a/litellm/utils.py +++ b/litellm/utils.py @@ -9010,7 +9010,7 @@ class ProviderConfigManager: return get_dashscope_family_rerank_config(provider.value) from litellm.llms.greenpt.rerank.transformation import GreenPTRerankConfig - config_class = { + config_class: Final = { litellm.LlmProviders.GREENPT: GreenPTRerankConfig, }.get(provider, litellm.CohereRerankConfig) return config_class() diff --git a/tests/unit/llms/greenpt/rerank/test_greenpt_rerank_transformation.py b/tests/unit/llms/greenpt/rerank/test_greenpt_rerank_transformation.py index 5eb6c7c0c7a..3153a4ce9db 100644 --- a/tests/unit/llms/greenpt/rerank/test_greenpt_rerank_transformation.py +++ b/tests/unit/llms/greenpt/rerank/test_greenpt_rerank_transformation.py @@ -1,6 +1,3 @@ -import json -from pathlib import Path - import pytest import litellm @@ -8,19 +5,6 @@ from litellm.cost_calculator import rerank_cost from litellm.llms.greenpt.rerank.transformation import GreenPTRerankConfig -@pytest.fixture -def local_model_cost_map(): - original_model_cost = litellm.model_cost - model_cost_path = Path(__file__).resolve().parents[5] / "model_prices_and_context_window.json" - litellm.model_cost = json.loads(model_cost_path.read_text(encoding="utf-8")) - litellm.get_model_info.cache_clear() - try: - yield - finally: - litellm.model_cost = original_model_cost - litellm.get_model_info.cache_clear() - - def test_greenpt_supported_params(): assert GreenPTRerankConfig().get_supported_cohere_rerank_params("green-rerank") == [ "query", @@ -94,7 +78,7 @@ def test_calculate_rerank_cost(): assert completion_cost == 0.0 -def test_public_rerank_cost_uses_token_pricing(local_model_cost_map): +def test_public_rerank_cost_uses_token_pricing(): prompt_cost, completion_cost = rerank_cost( model="greenpt/green-rerank", custom_llm_provider=None, diff --git a/tests/unit/llms/openai_like/test_greenpt_provider.py b/tests/unit/llms/openai_like/test_greenpt_provider.py index bf78b5e94be..dd8595c98f1 100644 --- a/tests/unit/llms/openai_like/test_greenpt_provider.py +++ b/tests/unit/llms/openai_like/test_greenpt_provider.py @@ -1,10 +1,6 @@ import json -import os -import sys from unittest.mock import Mock, patch -sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), "../../../.."))) - import litellm from litellm.llms.custom_httpx.http_handler import HTTPHandler from litellm.llms.openai_like.dynamic_config import create_config_class From 8863bd4f98306d0e8c1c13aa3a30d6157c2455c1 Mon Sep 17 00:00:00 2001 From: Robert Keus Date: Sun, 11 Oct 2026 08:31:36 +0200 Subject: [PATCH 09/10] fix(greenpt): accept litellm_params in rerank validate_environment The rerank handler passes litellm_params to validate_environment, so the GreenPT override raised TypeError on every litellm.rerank call. Adds a test that drives litellm.rerank through the HTTP handler, types the override so the basedpyright gate stays at its base count, and fills the interactions endpoint key. Co-Authored-By: Claude Opus 5.5 --- litellm/llms/greenpt/rerank/transformation.py | 37 ++++++++++--------- litellm/llms/openai_like/json_loader.py | 15 ++++---- provider_endpoints_support.json | 3 +- .../test_greenpt_rerank_transformation.py | 29 +++++++++++++++ 4 files changed, 59 insertions(+), 25 deletions(-) diff --git a/litellm/llms/greenpt/rerank/transformation.py b/litellm/llms/greenpt/rerank/transformation.py index d62c5d5538f..01b97bb300c 100644 --- a/litellm/llms/greenpt/rerank/transformation.py +++ b/litellm/llms/greenpt/rerank/transformation.py @@ -1,3 +1,4 @@ +from collections.abc import Mapping, Sized from typing import Final from litellm.llms.hosted_vllm.rerank.transformation import HostedVLLMRerankConfig @@ -7,41 +8,43 @@ from litellm.types.utils import ModelInfo class GreenPTRerankConfig(HostedVLLMRerankConfig): - def get_supported_cohere_rerank_params(self, model: str) -> list: # mutable-ok: mirrors override contract + def get_supported_cohere_rerank_params(self, model: str) -> list[str]: # mutable-ok: mirrors override contract return ["query", "documents", "top_n", "return_documents"] def validate_environment( self, - headers: dict, # mutable-ok: mirrors override contract + headers: dict[str, str], # mutable-ok: mirrors override contract model: str, api_key: str | None = None, - optional_params: dict | None = None, # mutable-ok: mirrors override contract - ) -> dict: # mutable-ok: mirrors override contract + optional_params: dict[str, object] | None = None, # mutable-ok: mirrors override contract + litellm_params: Mapping[str, object] | None = None, + ) -> dict[str, str]: # mutable-ok: mirrors override contract resolved_api_key: Final = api_key or get_secret_str("GREENPT_API_KEY") if resolved_api_key is None: raise ValueError("GreenPT API key is required. Set GREENPT_API_KEY.") - return super().validate_environment( - headers=headers, - model=model, - api_key=resolved_api_key, - optional_params=optional_params, - ) + return { + "Authorization": f"Bearer {resolved_api_key}", + "accept": "application/json", + "content-type": "application/json", + **headers, + } def transform_rerank_request( self, model: str, - optional_rerank_params: dict, # mutable-ok: mirrors override contract - headers: dict, # mutable-ok: mirrors override contract - litellm_params: dict | None = None, # mutable-ok: mirrors override contract - ) -> dict: # mutable-ok: mirrors override contract - request: Final = super().transform_rerank_request( + optional_rerank_params: dict[str, object], # mutable-ok: mirrors override contract + headers: dict[str, str], # mutable-ok: mirrors override contract + litellm_params: dict[str, object] | None = None, # mutable-ok: mirrors override contract + ) -> dict[str, object]: # mutable-ok: mirrors override contract + request: Final[Mapping[str, object]] = super().transform_rerank_request( # pyright: ignore[reportUnknownMemberType, reportUnknownVariableType] # HostedVLLMRerankConfig returns an untyped dict model=model, optional_rerank_params=optional_rerank_params, headers=headers, litellm_params=litellm_params, ) - request.setdefault("top_n", len(request["documents"])) - return request + documents: Final = optional_rerank_params["documents"] + top_n_default: Final = len(documents) if isinstance(documents, Sized) else None + return {"top_n": top_n_default, **request} if top_n_default is not None else {**request} def calculate_rerank_cost( self, diff --git a/litellm/llms/openai_like/json_loader.py b/litellm/llms/openai_like/json_loader.py index 7bbe523b0c0..d922399d1f8 100644 --- a/litellm/llms/openai_like/json_loader.py +++ b/litellm/llms/openai_like/json_loader.py @@ -71,19 +71,20 @@ class JSONProviderRegistry: return next((provider for provider in cls._providers.values() if provider.base_url == base_url), None) @classmethod - def supports_responses_api(cls, slug: str) -> bool: - """Check if a JSON provider supports the Responses API""" + def _supports_endpoint(cls, slug: str, endpoint: str) -> bool: provider: Final = cls._providers.get(slug) if provider is None: return False - return "/v1/responses" in provider.supported_endpoints + return endpoint in provider.supported_endpoints + + @classmethod + def supports_responses_api(cls, slug: str) -> bool: + """Check if a JSON provider supports the Responses API""" + return cls._supports_endpoint(slug, "/v1/responses") @classmethod def supports_embeddings_api(cls, slug: str) -> bool: - provider: Final = cls._providers.get(slug) - if provider is None: - return False - return "/v1/embeddings" in provider.supported_endpoints + return cls._supports_endpoint(slug, "/v1/embeddings") @classmethod def list_providers(cls) -> list: diff --git a/provider_endpoints_support.json b/provider_endpoints_support.json index e4c92e2eba9..1cfc9888a44 100644 --- a/provider_endpoints_support.json +++ b/provider_endpoints_support.json @@ -3221,7 +3221,8 @@ "moderations": false, "batches": false, "rerank": true, - "a2a": false + "a2a": false, + "interactions": false } } }, diff --git a/tests/unit/llms/greenpt/rerank/test_greenpt_rerank_transformation.py b/tests/unit/llms/greenpt/rerank/test_greenpt_rerank_transformation.py index 3153a4ce9db..4ab67945198 100644 --- a/tests/unit/llms/greenpt/rerank/test_greenpt_rerank_transformation.py +++ b/tests/unit/llms/greenpt/rerank/test_greenpt_rerank_transformation.py @@ -1,7 +1,11 @@ +import json +from unittest.mock import Mock, patch + import pytest import litellm from litellm.cost_calculator import rerank_cost +from litellm.llms.custom_httpx.http_handler import HTTPHandler from litellm.llms.greenpt.rerank.transformation import GreenPTRerankConfig @@ -102,3 +106,28 @@ def test_calculate_rerank_cost_missing_usage_or_pricing(billed_units, model_info billed_units=billed_units, model_info=model_info, ) == (0.0, 0.0) + + +def test_public_rerank_call_sends_request_and_costs_tokens(): + response = Mock() + response.raise_for_status = Mock() + response.json.return_value = { + "id": "rerank-1", + "results": [{"index": 0, "relevance_score": 0.9}], + "usage": {"total_tokens": 1000}, + } + + with patch.object(HTTPHandler, "post", return_value=response) as mock_post: + result = litellm.rerank( + model="greenpt/green-rerank", + query="low-carbon inference", + documents=["GreenPT", "Other provider"], + api_key="test-key", + client=HTTPHandler(), + ) + + assert mock_post.call_args.kwargs["url"] == "https://api.greenpt.ai/v1/rerank" + assert mock_post.call_args.kwargs["headers"]["Authorization"] == "Bearer test-key" + assert json.loads(mock_post.call_args.kwargs["data"])["top_n"] == 2 + assert result.results[0]["index"] == 0 + assert result._hidden_params["response_cost"] == pytest.approx(0.000136524) From 9ddde3c6db84c618b2ee10281ab19b9f4a91b3b8 Mon Sep 17 00:00:00 2001 From: Robert Keus Date: Sun, 11 Oct 2026 08:34:18 +0200 Subject: [PATCH 10/10] ci(greenpt): add GreenPT tests to the LLM provider shard Co-Authored-By: Claude Opus 5.5 --- .github/workflows/test-unit.yml | 1 + 1 file changed, 1 insertion(+) diff --git a/.github/workflows/test-unit.yml b/.github/workflows/test-unit.yml index 18df865c988..594e99d4a0b 100644 --- a/.github/workflows/test-unit.yml +++ b/.github/workflows/test-unit.yml @@ -272,6 +272,7 @@ jobs: tests/unit/llms/gigachat tests/unit/llms/github_copilot tests/unit/llms/gradient_ai + tests/unit/llms/greenpt tests/unit/llms/groq tests/unit/llms/heroku tests/unit/llms/hosted_vllm