fix: enable JSON-only provider routing + add xiaomi_mimo provider (#18291)

* fix: enable JSON-only provider routing + add xiaomi_mimo provider

* docs: add xiaomi_mimo provider documentation

docs/my-website/docs/providers/xiaomi_mimo.md

@@ -0,0 +1,137 @@
+import Tabs from '@theme/Tabs';
+import TabItem from '@theme/TabItem';
+
+# Xiaomi MiMo
+https://platform.xiaomimimo.com/#/docs
+
+:::tip
+
+**We support ALL Xiaomi MiMo models, just set `model=xiaomi_mimo/<any-model-on-xiaomi-mimo>` as a prefix when sending litellm requests**
+
+:::
+
+## API Key
+```python
+# env variable
+os.environ['XIAOMI_MIMO_API_KEY']
+```
+
+## Sample Usage
+```python
+from litellm import completion
+import os
+
+os.environ['XIAOMI_MIMO_API_KEY'] = ""
+response = completion(
+    model="xiaomi_mimo/mimo-v2-flash",
+    messages=[
+        {
+            "role": "user",
+            "content": "What's the weather like in Boston today in Fahrenheit?",
+        }
+    ],
+    max_tokens=1024,
+    temperature=0.3,
+    top_p=0.95,
+)
+print(response)
+```
+
+## Sample Usage - Streaming
+```python
+from litellm import completion
+import os
+
+os.environ['XIAOMI_MIMO_API_KEY'] = ""
+response = completion(
+    model="xiaomi_mimo/mimo-v2-flash",
+    messages=[
+        {
+            "role": "user",
+            "content": "What's the weather like in Boston today in Fahrenheit?",
+        }
+    ],
+    stream=True,
+    max_tokens=1024,
+    temperature=0.3,
+    top_p=0.95,
+)
+
+for chunk in response:
+    print(chunk)
+```
+
+
+## Usage with LiteLLM Proxy Server
+
+Here's how to call a Xiaomi MiMo model with the LiteLLM Proxy Server
+
+1. Modify the config.yaml 
+
+  ```yaml
+  model_list:
+    - model_name: my-model
+      litellm_params:
+        model: xiaomi_mimo/<your-model-name>  # add xiaomi_mimo/ prefix to route as Xiaomi MiMo provider
+        api_key: api-key                      # api key to send your model
+  ```
+
+
+2. Start the proxy 
+
+  ```bash
+  $ litellm --config /path/to/config.yaml
+  ```
+
+3. Send Request to LiteLLM Proxy Server
+
+  <Tabs>
+
+  <TabItem value="openai" label="OpenAI Python v1.0.0+">
+
+  ```python
+  import openai
+  client = openai.OpenAI(
+      api_key="sk-1234",             # pass litellm proxy key, if you're using virtual keys
+      base_url="http://0.0.0.0:4000" # litellm-proxy-base url
+  )
+
+  response = client.chat.completions.create(
+      model="my-model",
+      messages = [
+          {
+              "role": "user",
+              "content": "what llm are you"
+          }
+      ],
+  )
+
+  print(response)
+  ```
+  </TabItem>
+
+  <TabItem value="curl" label="curl">
+
+  ```shell
+  curl --location 'http://0.0.0.0:4000/chat/completions' \
+      --header 'Authorization: Bearer sk-1234' \
+      --header 'Content-Type: application/json' \
+      --data '{
+      "model": "my-model",
+      "messages": [
+          {
+          "role": "user",
+          "content": "what llm are you"
+          }
+      ],
+  }'
+  ```
+  </TabItem>
+
+  </Tabs>
+
+## Supported Models
+
+| Model Name | Usage |
+|------------|-------|
+| mimo-v2-flash | `completion(model="xiaomi_mimo/mimo-v2-flash", messages)` |

docs/my-website/sidebars.js

@@ -289,7 +289,7 @@ const sidebars = {
           label: "All Endpoints (Swagger)",
           href: "https://litellm-api.up.railway.app/",
         },
-  "proxy/enterprise",
+        "proxy/enterprise",
         {
           type: "category",
           label: "Authentication",
@@ -470,10 +470,10 @@ const sidebars = {
             "proxy/managed_finetuning",
           ]
         },
-          "generateContent",
-          "apply_guardrail",
-          "bedrock_invoke",
-          "interactions",
+        "generateContent",
+        "apply_guardrail",
+        "bedrock_invoke",
+        "interactions",
         {
           type: "category",
           label: "/images",
@@ -783,6 +783,7 @@ const sidebars = {
           ]
         },
         "providers/xai",
+        "providers/xiaomi_mimo",
         "providers/xinference",
         "providers/zai",
       ],

litellm/litellm_core_utils/get_llm_provider_logic.py

@@ -5,6 +5,7 @@ import httpx
 import litellm
 from litellm.constants import REPLICATE_MODEL_NAME_WITH_ID_LENGTH
 from litellm.secret_managers.main import get_secret, get_secret_str
+from litellm.llms.openai_like.json_loader import JSONProviderRegistry
 
 from ..types.router import LiteLLM_Params
 
@@ -155,6 +156,17 @@ def get_llm_provider(  # noqa: PLR0915
 
         if api_key and api_key.startswith("os.environ/"):
             dynamic_api_key = get_secret_str(api_key)
+
+        # Check JSON-configured providers FIRST (before enum-based provider_list)
+        provider_prefix = model.split("/", 1)[0]
+        if len(model.split("/")) > 1 and JSONProviderRegistry.exists(provider_prefix):
+            return _get_openai_compatible_provider_info(
+                model=model,
+                api_base=api_base,
+                api_key=api_key,
+                dynamic_api_key=dynamic_api_key,
+            )
+
         # check if llm provider part of model name
 
         if (

litellm/llms/openai/openai.py

@@ -555,9 +555,13 @@ class OpenAIChatCompletion(BaseLLM, BaseOpenAILLM):
             provider_config: Optional[BaseConfig] = None
 
             if custom_llm_provider is not None and model is not None:
-                provider_config = ProviderConfigManager.get_provider_chat_config(
-                    model=model, provider=LlmProviders(custom_llm_provider)
-                )
+                try:
+                    provider_config = ProviderConfigManager.get_provider_chat_config(
+                        model=model, provider=LlmProviders(custom_llm_provider)
+                    )
+                except ValueError:
+                    # JSON-configured providers may not be in LlmProviders enum
+                    provider_config = None
 
             if provider_config is None:
                 provider_config = OpenAIConfig()

litellm/llms/openai_like/providers.json

@@ -18,5 +18,12 @@
   "veniceai": {
     "base_url": "https://api.venice.ai/api/v1",
     "api_key_env": "VENICE_AI_API_KEY"
+  },
+  "xiaomi_mimo": {
+    "base_url": "https://api.xiaomimimo.com/v1",
+    "api_key_env": "XIAOMI_MIMO_API_KEY",
+    "param_mappings": {
+      "max_completion_tokens": "max_tokens"
+    }
   }
-}
+}

litellm/main.py

@@ -68,6 +68,7 @@ from litellm.constants import (
     DEFAULT_MOCK_RESPONSE_PROMPT_TOKEN_COUNT,
 )
 from litellm.exceptions import LiteLLMUnknownProvider
+from litellm.llms.openai_like.json_loader import JSONProviderRegistry
 from litellm.integrations.custom_logger import CustomLogger
 from litellm.litellm_core_utils.asyncify import run_async_function
 from litellm.litellm_core_utils.audio_utils.utils import (
@@ -2263,6 +2264,7 @@ def completion(  # type: ignore # noqa: PLR0915
             or custom_llm_provider == "wandb"
             or custom_llm_provider == "clarifai"
             or custom_llm_provider in litellm.openai_compatible_providers
+            or JSONProviderRegistry.exists(custom_llm_provider)  # JSON-configured providers
             or "ft:gpt-3.5-turbo" in model  # finetune gpt-3.5-turbo
         ):  # allow user to make an openai call with a custom base
             # note: if a user sets a custom base - we should ensure this works