{"type":"post","id":"google-open-sources-xla-serving-engine-for-train-9244ff","slug":"google-open-sources-xla-serving-engine-for-train-9244ff","title":"Google open-sources XLA Serving Engine for train-in-Python, serve-in-C++ inference","description":"Google has open-sourced XLA Serving Engine (XSE), an Apache 2.0 inference runtime that lets teams export models from Python into an XLA SavedModel and serve them in C++.","retrieval_nugget":"Google has open-sourced XLA Serving Engine (XSE), an Apache 2.0 inference runtime that lets teams export models from Python into an XLA SavedModel and serve them in C++. The release matters to teams weighing serving latency and operational complexity: Google reports substantial gains on short requests, while explicitly steering long-context autoregressive LLM workloads toward other engines.","published_at":"2026-11-05","updated_at":"2026-11-05","record_date":"2026-11-05","date_kind":"published_at","topics":["google"],"entities":["Google","XLA Serving Engine","OpenXLA","JAX","PyTorch","Arm","Intel","Moloco"],"related_patterns":[],"novelty_label":"new_fact","what_changed":"Google is releasing a distinct XLA Serving Engine for exporting models to run in C++, reporting up to 6 million queries per second internally and performance gains over Python serving.","prior_refs":["https://newruntime.com/posts/announcing-new-vllm-and-sglang-inference-engines-on-be000f/","https://newruntime.com/posts/announcing-llm-d-1-0-open-distributed-inference-7539eb/"],"source_url":"https://opensource.googleblog.com/2026/11/supercharging-xla-inference-xla-serving-engine.html","source_urls":["https://opensource.googleblog.com/2026/11/supercharging-xla-inference-xla-serving-engine.html"],"schema_version":"newruntime-agent-readable-v0.2","stable_id":"post:google-open-sources-xla-serving-engine-for-train-9244ff","status":"published","visuals":[],"editor":{"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]},"routes":{"html":"https://newruntime.com/posts/google-open-sources-xla-serving-engine-for-train-9244ff/","markdown":"https://newruntime.com/posts/google-open-sources-xla-serving-engine-for-train-9244ff.md","json":"https://newruntime.com/posts/google-open-sources-xla-serving-engine-for-train-9244ff.json"},"markdown_url":"https://newruntime.com/posts/google-open-sources-xla-serving-engine-for-train-9244ff.md","next_reads":[{"type":"topic","path":"/topics/google/","url":"https://newruntime.com/topics/google/","title":"Google","reason":"Explore the google topic hub.","media_type":"text/html","markdown_url":"https://newruntime.com/topics/google.md","json_url":"https://newruntime.com/topics/google.json"},{"type":"collection","path":"/posts/","url":"https://newruntime.com/posts/","title":"Field Notes","reason":"Open the latest editorial analysis.","media_type":"text/html","json_url":"https://newruntime.com/posts.json"},{"type":"collection","path":"/sources/","url":"https://newruntime.com/sources/","title":"Sources","reason":"See the original sources this archive cites.","media_type":"text/html"}]}