feat(proxy_server.py): expose new /model_group/info endpoint

returns model-group level info on supported params, max tokens, pricing, etc.

litellm/proxy/auth/model_checks.py

@@ -60,17 +60,20 @@ def get_complete_model_list(
     - If team list is empty -> defer to proxy model list
     """
 
-    if len(key_models) > 0:
+    unique_models = set()
-        return key_models
 
-    if len(team_models) > 0:
+    if key_models:
-        return team_models
+        unique_models.update(key_models)
+    elif team_models:
+        unique_models.update(team_models)
+    else:
+        unique_models.update(proxy_model_list)
 
-    returned_models = proxy_model_list
+        if user_model:
-    if user_model is not None:  # set via `litellm --model ollama/llam3`
+            unique_models.add(user_model)
-        returned_models.append(user_model)
 
-    if infer_model_from_keys is not None and infer_model_from_keys == True:
+        if infer_model_from_keys:
             valid_models = get_valid_models()
-        returned_models.extend(valid_models)
+            unique_models.update(valid_models)
-    return returned_models
+
+    return list(unique_models)

litellm/proxy/proxy_server.py

@@ -2,7 +2,7 @@ import sys, os, platform, time, copy, re, asyncio, inspect
 import threading, ast
 import shutil, random, traceback, requests
 from datetime import datetime, timedelta, timezone
-from typing import Optional, List, Callable, get_args
+from typing import Optional, List, Callable, get_args, Set
 import secrets, subprocess
 import hashlib, uuid
 import warnings
@@ -106,7 +106,7 @@ import pydantic
 from litellm.proxy._types import *
 from litellm.caching import DualCache, RedisCache
 from litellm.proxy.health_check import perform_health_check
-from litellm.router import LiteLLM_Params, Deployment, updateDeployment
+from litellm.router import LiteLLM_Params, Deployment, updateDeployment, ModelGroupInfo
 from litellm.router import ModelInfo as RouterModelInfo
 from litellm._logging import verbose_router_logger, verbose_proxy_logger
 from litellm.proxy.auth.handle_jwt import JWTHandler
@@ -9730,6 +9730,58 @@ async def model_info_v1(
     return {"data": all_models}
 
 
+@router.get(
+    "/model_group/info",
+    description="Provides more info about each model in /models, including config.yaml descriptions (except api key and api base)",
+    tags=["model management"],
+    dependencies=[Depends(user_api_key_auth)],
+)
+async def model_group_info(
+    user_api_key_dict: UserAPIKeyAuth = Depends(user_api_key_auth),
+):
+    """
+    Returns model info at the model group level.
+    """
+    global llm_model_list, general_settings, user_config_file_path, proxy_config, llm_router
+
+    if llm_model_list is None:
+        raise HTTPException(
+            status_code=500, detail={"error": "LLM Model List not loaded in"}
+        )
+    if llm_router is None:
+        raise HTTPException(
+            status_code=500, detail={"error": "LLM Router is not loaded in"}
+        )
+    all_models: List[dict] = []
+    ## CHECK IF MODEL RESTRICTIONS ARE SET AT KEY/TEAM LEVEL ##
+    if llm_model_list is None:
+        proxy_model_list = []
+    else:
+        proxy_model_list = [m["model_name"] for m in llm_model_list]
+    key_models = get_key_models(
+        user_api_key_dict=user_api_key_dict, proxy_model_list=proxy_model_list
+    )
+    team_models = get_team_models(
+        user_api_key_dict=user_api_key_dict, proxy_model_list=proxy_model_list
+    )
+    all_models_str = get_complete_model_list(
+        key_models=key_models,
+        team_models=team_models,
+        proxy_model_list=proxy_model_list,
+        user_model=user_model,
+        infer_model_from_keys=general_settings.get("infer_model_from_keys", False),
+    )
+
+    model_groups: List[ModelGroupInfo] = []
+    for model in all_models_str:
+
+        _model_group_info = llm_router.get_model_group_info(model_group=model)
+        if _model_group_info is not None:
+            model_groups.append(_model_group_info)
+
+    return {"data": model_groups}
+
+
 #### [BETA] - This is a beta endpoint, format might change based on user feedback. - https://github.com/BerriAI/litellm/issues/964
 @router.post(
     "/model/delete",

litellm/router.py

@@ -48,6 +48,7 @@ from litellm.types.router import (
     RetryPolicy,
     AlertingConfig,
     DeploymentTypedDict,
+    ModelGroupInfo,
 )
 from litellm.integrations.custom_logger import CustomLogger
 from litellm.llms.azure import get_azure_ad_token_from_oidc
@@ -3045,6 +3046,100 @@ class Router:
                     return model
         return None
 
+    def get_model_group_info(self, model_group: str) -> Optional[ModelGroupInfo]:
+        """
+        For a given model group name, return the combined model info
+
+        Returns:
+        - ModelGroupInfo if able to construct a model group
+        - None if error constructing model group info
+        """
+
+        model_group_info: Optional[ModelGroupInfo] = None
+
+        for model in self.model_list:
+            if "model_name" in model and model["model_name"] == model_group:
+                # model in model group found #
+                litellm_params = LiteLLM_Params(**model["litellm_params"])
+                # get model info
+                try:
+                    model_info = litellm.get_model_info(model=litellm_params.model)
+                except Exception as e:
+                    continue
+                # get llm provider
+                try:
+                    model, llm_provider, _, _ = litellm.get_llm_provider(
+                        model=litellm_params.model,
+                        custom_llm_provider=litellm_params.custom_llm_provider,
+                    )
+                except Exception as e:
+                    continue
+
+                if model_group_info is None:
+                    model_group_info = ModelGroupInfo(
+                        model_group=model_group, providers=[llm_provider], **model_info  # type: ignore
+                    )
+                else:
+                    # if max_input_tokens > curr
+                    # if max_output_tokens > curr
+                    # if input_cost_per_token > curr
+                    # if output_cost_per_token > curr
+                    # supports_parallel_function_calling == True
+                    # supports_vision == True
+                    # supports_function_calling == True
+                    if llm_provider not in model_group_info.providers:
+                        model_group_info.providers.append(llm_provider)
+                    if model_info.get("max_input_tokens", None) is not None and (
+                        model_group_info.max_input_tokens is None
+                        or model_info["max_input_tokens"]
+                        > model_group_info.max_input_tokens
+                    ):
+                        model_group_info.max_input_tokens = model_info[
+                            "max_input_tokens"
+                        ]
+                    if model_info.get("max_output_tokens", None) is not None and (
+                        model_group_info.max_output_tokens is None
+                        or model_info["max_output_tokens"]
+                        > model_group_info.max_output_tokens
+                    ):
+                        model_group_info.max_output_tokens = model_info[
+                            "max_output_tokens"
+                        ]
+                    if model_info.get("input_cost_per_token", None) is not None and (
+                        model_group_info.input_cost_per_token is None
+                        or model_info["input_cost_per_token"]
+                        > model_group_info.input_cost_per_token
+                    ):
+                        model_group_info.input_cost_per_token = model_info[
+                            "input_cost_per_token"
+                        ]
+                    if model_info.get("output_cost_per_token", None) is not None and (
+                        model_group_info.output_cost_per_token is None
+                        or model_info["output_cost_per_token"]
+                        > model_group_info.output_cost_per_token
+                    ):
+                        model_group_info.output_cost_per_token = model_info[
+                            "output_cost_per_token"
+                        ]
+                    if (
+                        model_info.get("supports_parallel_function_calling", None)
+                        is not None
+                        and model_info["supports_parallel_function_calling"] == True  # type: ignore
+                    ):
+                        model_group_info.supports_parallel_function_calling = True
+                    if (
+                        model_info.get("supports_vision", None) is not None
+                        and model_info["supports_vision"] == True  # type: ignore
+                    ):
+                        model_group_info.supports_vision = True
+                    if (
+                        model_info.get("supports_function_calling", None) is not None
+                        and model_info["supports_function_calling"] == True  # type: ignore
+                    ):
+                        model_group_info.supports_function_calling = True
+
+        return model_group_info
+
     def get_model_ids(self) -> List[str]:
         """
         Returns list of model id's.

litellm/types/router.py

@@ -411,3 +411,18 @@ class AlertingConfig(BaseModel):
 
     webhook_url: str
     alerting_threshold: Optional[float] = 300
+
+
+class ModelGroupInfo(BaseModel):
+    model_group: str
+    providers: List[str]
+    max_input_tokens: Optional[float] = None
+    max_output_tokens: Optional[float] = None
+    input_cost_per_token: Optional[float] = None
+    output_cost_per_token: Optional[float] = None
+    mode: Literal[
+        "chat", "embedding", "completion", "image_generation", "audio_transcription"
+    ]
+    supports_parallel_function_calling: bool = Field(default=False)
+    supports_vision: bool = Field(default=False)
+    supports_function_calling: bool = Field(default=False)

litellm/types/utils.py

@@ -12,3 +12,13 @@ class ProviderField(TypedDict):
     field_type: Literal["string"]
     field_description: str
     field_value: str
+
+
+class ModelInfo(TypedDict):
+    max_tokens: int
+    max_input_tokens: int
+    max_output_tokens: int
+    input_cost_per_token: float
+    output_cost_per_token: float
+    litellm_provider: str
+    mode: str

litellm/utils.py

@@ -34,7 +34,7 @@ from dataclasses import (
 import litellm._service_logger  # for storing API inputs, outputs, and metadata
 from litellm.llms.custom_httpx.http_handler import HTTPHandler
 from litellm.caching import DualCache
-from litellm.types.utils import CostPerToken, ProviderField
+from litellm.types.utils import CostPerToken, ProviderField, ModelInfo
 
 oidc_cache = DualCache()
 
@@ -7092,7 +7092,7 @@ def get_max_tokens(model: str):
         )
 
 
-def get_model_info(model: str):
+def get_model_info(model: str) -> ModelInfo:
     """
     Get a dict for the maximum tokens (context window),
     input_cost_per_token, output_cost_per_token  for a given model.
@@ -7154,7 +7154,7 @@ def get_model_info(model: str):
         if custom_llm_provider == "huggingface":
             max_tokens = _get_max_position_embeddings(model_name=model)
             return {
-                "max_tokens": max_tokens,
+                "max_tokens": max_tokens,  # type: ignore
                 "input_cost_per_token": 0,
                 "output_cost_per_token": 0,
                 "litellm_provider": "huggingface",