fix(rust): box messages response output

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>

litellm-rust/crates/core/src/messages/mod.rs

@@ -34,7 +34,7 @@ pub async fn messages(request: MessagesRequest<'_>) -> Result<AnthropicMessagesR
         timeout: request.timeout,
     };
     match litellm_host::run::run(messages_machine(), &LocalMessagesHost::new(call)).await? {
-        MessagesOutput::Message(message) => Ok(message),
+        MessagesOutput::Message(message) => Ok(*message),
         MessagesOutput::Streamed => Err(Error::Unsupported(
             "streamed responses need a streaming host",
         )),

litellm-rust/crates/core/src/messages/route.rs

@@ -48,7 +48,7 @@ impl MessagesCall {
 }
 
 pub enum MessagesOutput {
-    Message(AnthropicMessagesResponse),
+    Message(Box<AnthropicMessagesResponse>),
     /// Every chunk already reached the host through `Deliver`.
     Streamed,
 }
@@ -174,7 +174,8 @@ async fn execute(host: MessagesHost) -> Result<MessagesOutput, Error> {
         raw: RawResponse { body: text.clone() },
     })
     .await?;
-    decode_response(request.config, &request.model, &text).map(MessagesOutput::Message)
+    decode_response(request.config, &request.model, &text)
+        .map(|message| MessagesOutput::Message(Box::new(message)))
 }
 
 /// Hands each upstream chunk to the caller as it arrives. A caller that stops reading

litellm-rust/crates/python-bridge/src/routes/messages/host.rs

@@ -150,7 +150,7 @@ impl RouteHost for MessagesRouteHost {
             MessagesOutput::Message(message) => py
                 .import("litellm.rust_bridge.messages.route_host")?
                 .getattr("response")?
-                .call1((to_py(py, &message)?,))
+                .call1((to_py(py, message.as_ref())?,))
                 .map(Bound::unbind),
             MessagesOutput::Streamed => Ok(py.None()),
         }