server: add gpt-oss chat template for proper prompt formatting

The gpt-oss model requires a specific prompt format with <|start|>,
<|message|>, <|end|>, <|channel|> tokens. Without this, the model
produces degenerate output. Auto-detected via config.model_type.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

crates/xserv-server/src/api.rs

@@ -89,7 +89,7 @@ async fn chat_non_stream(state: Arc<AppState>, req: ChatRequest) -> Response {
         return response;
     }
 
-    let prompt = build_prompt(&req.messages);
+    let prompt = build_prompt(&req.messages, &state.model_type);
     let prompt_tokens = state.engine_tokenizer.lock().unwrap().encode(&prompt);
     let prompt_token_count = prompt_tokens.len();
 
@@ -159,7 +159,7 @@ fn chat_stream(
         return response;
     }
 
-    let prompt = build_prompt(&req.messages);
+    let prompt = build_prompt(&req.messages, &state.model_type);
     let prompt_tokens = state.engine_tokenizer.lock().unwrap().encode(&prompt);
 
     let max_seq_len = state.max_seq_len;
@@ -325,7 +325,11 @@ fn sampling_params(req: &ChatRequest) -> SamplingParams {
     }
 }
 
-fn build_prompt(messages: &[Message]) -> String {
+fn build_prompt(messages: &[Message], model_type: &str) -> String {
+    if model_type == "gpt_oss" {
+        return build_prompt_gpt_oss(messages);
+    }
+    // Default: Qwen3 ChatML format
     let mut prompt = String::new();
     for msg in messages {
         match msg.role.as_str() {
@@ -343,3 +347,28 @@ fn build_prompt(messages: &[Message]) -> String {
     prompt.push_str("<think>\n\n</think>\n\n");
     prompt
 }
+
+fn build_prompt_gpt_oss(messages: &[Message]) -> String {
+    let mut prompt = String::new();
+    // System prompt
+    prompt.push_str("<|start|>system<|message|>");
+    prompt.push_str("You are a helpful assistant.\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.");
+    prompt.push_str("<|end|>");
+    for msg in messages {
+        match msg.role.as_str() {
+            "user" => {
+                prompt.push_str("<|start|>user<|message|>");
+                prompt.push_str(&msg.content);
+                prompt.push_str("<|end|>");
+            }
+            "assistant" => {
+                prompt.push_str("<|start|>assistant<|channel|>final<|message|>");
+                prompt.push_str(&msg.content);
+                prompt.push_str("<|end|>");
+            }
+            _ => {}
+        }
+    }
+    prompt.push_str("<|start|>assistant");
+    prompt
+}

crates/xserv-server/src/main.rs

@@ -11,6 +11,7 @@ use xserv_model::ModelConfig;
 
 pub struct AppState {
     pub model_name: String,
+    pub model_type: String,
     pub engine_sender: Mutex<mpsc::Sender<GenerateRequest>>,
     pub engine_tokenizer: Mutex<xserv_tokenizer::Tokenizer>,
     pub max_seq_len: usize,
@@ -99,8 +100,10 @@ async fn main() {
         }
     });
 
+    let model_type = model_config.model_type.clone().unwrap_or_default();
     let state = Arc::new(AppState {
         model_name,
+        model_type,
         engine_sender: Mutex::new(tx),
         engine_tokenizer: Mutex::new(tokenizer),
         max_seq_len,