fixed token count reading

packages/inference-service/src/providers/llamacpp.js

@@ -26,6 +26,7 @@ function buildPayload(prompt, options, stream = false) {
     top_k: opts.topK,
     repeat_penalty: opts.repeatPenalty,
     stream,
+    stream_options: stream ? { include_usage: true } : undefined,
     ...(opts.seed !== null && { seed: opts.seed }),
   };
 }