Update the openai completion request body

app/api/schemas/openai.py

@@ -41,33 +41,36 @@ class ChatMessage(BaseModel):
 
 
 class ChatCompletionRequest(BaseModel):
-    model: str
     messages: list[ChatMessage]
-    temperature: float | None = 1.0
-    top_p: float | None = 1.0
-    n: int | None = 1
-    stream: bool | None = False
-    stream_options: dict[str, Any] | None = None
-    stop: str | list[str] | None = None
-    max_tokens: int | None = None
-    presence_penalty: float | None = 0.0
+    model: str
+    audio: object | None = None
     frequency_penalty: float | None = 0.0
-    logit_bias: dict[str, float] | None = None
-    user: str | None = None
-    web_search_options: Any | None = None
-    response_format: dict[str, Any] | None = None
-    tools: list[Any, Any] | None = None
-    tool_choice: str | None = None
-    seed: int | None = None
+    logit_bias: dict[Any, Any] | None = None
     logprobs: bool | None = None
-    top_logprobs: int | None = None
     max_completion_tokens: int | None = None
+    max_tokens: int | None = None # deprecated
     metadata: dict[Any, Any] | None = None
-    modalities: list[str] | None = None
+    modalities: list[Any] | None = None
+    n: int | None = 1
     parallel_tool_calls: bool | None = True
-    prediction: dict[Any, Any] | None = None
+    prediction: object | None = None
+    presence_penalty: float | None = 0.0
     reasoning_effort: str | None = "medium"
+    response_format: Any | None = None
+    seed: int | None = None
     service_tier: str | None = "auto"
+    stop: str | list[Any] | None = None
+    store: bool | None = False
+    stream: bool | None = False
+    stream_options: object | None = None
+    temperature: float | None = 1.0
+    tool_choice: str | None = None
+    tools: list[Any] | None = None
+    top_logprobs: int | None = None
+    top_p: float | None = 1.0
+    user: str | None = None
+    web_search_options: object | None = None
+
 
 class ChatCompletionResponseChoice(BaseModel):
     index: int
@@ -93,19 +96,20 @@ class ChatCompletionResponse(BaseModel):
 class CompletionRequest(BaseModel):
     model: str
     prompt: str | list[str]
-    suffix: str | None = None
+    best_of: int | None = 1
+    echo: bool | None = False
+    frequency_penalty: float | None = 0.0
+    logit_bias: dict[Any, Any] | None = None
+    logprobs: int | None = None
     max_tokens: int | None = 16
-    temperature: float | None = 1.0
-    top_p: float | None = 1.0
     n: int | None = 1
-    stream: bool | None = False
-    logprobs: int | None = None
-    echo: bool | None = False
-    stop: str | list[str] | None = None
     presence_penalty: float | None = 0.0
-    frequency_penalty: float | None = 0.0
-    best_of: int | None = 1
-    logit_bias: dict[str, float] | None = None
+    seed: int | None = None
+    stop: str | list[str] | None = None
+    stream: bool | None = False
+    stream_options: object | None = None
+    suffix: str | None = None
+    temperature: float | None = 1.0
     user: str | None = None
 
 

app/services/providers/anthropic_adapter.py

@@ -124,7 +124,7 @@ async def process_completion(
         streaming = payload.get("stream", False)
         anthropic_payload = {
             "model": payload["model"],
-            "max_tokens": payload.get("max_tokens", ANTHROPIC_DEFAULT_MAX_TOKENS),
+            "max_tokens": payload.get("max_completion_tokens", payload.get("max_tokens", ANTHROPIC_DEFAULT_MAX_TOKENS)),
             "temperature": payload.get("temperature", 1.0),
             "stop_sequences": payload.get("stop", []),
         }

app/services/providers/bedrock_adapter.py

@@ -184,8 +184,8 @@ async def convert_openai_payload_to_bedrock(payload: dict[str, Any]) -> dict[str
         inferenceConfig = {}
         if "temperature" in payload:
             inferenceConfig["temperature"] = payload["temperature"]
-        if "max_tokens" in payload:
-            inferenceConfig["maxTokens"] = payload["max_tokens"]
+        if "max_completion_tokens" in payload or "max_tokens" in payload:
+            inferenceConfig["maxTokens"] = payload.get("max_completion_tokens", payload.get("max_tokens"))
         if "top_p" in payload:
             inferenceConfig["topP"] = payload["top_p"]
         if "stop" in payload:

app/services/providers/google_adapter.py

@@ -392,7 +392,7 @@ async def convert_openai_completion_payload_to_google(
                 "stopSequences": payload.get("stop", []),
                 "temperature": payload.get("temperature", 0.7),
                 "topP": payload.get("top_p", 0.95),
-                "maxOutputTokens": payload.get("max_tokens", 2048),
+                "maxOutputTokens": payload.get("max_completion_tokens", payload.get("max_tokens", 2048)),
             },
         }
 

app/services/providers/mock_adapter.py

@@ -52,7 +52,7 @@ async def process_completion(
             messages = payload.get("messages", [])
             temperature = payload.get("temperature", 0.7)
             stream = payload.get("stream", False)
-            max_tokens = payload.get("max_tokens")
+            max_tokens = payload.get("max_completion_tokens", payload.get("max_tokens"))
 
             if stream:
                 # For streaming, return a generator