{"type":"post","id":"quark-our-whole-training-stack-end-to-end-eba988","slug":"quark-our-whole-training-stack-end-to-end-eba988","title":"Quark: our whole training stack, end to end, for $4,200","description":"Empero says it trained and released Quark, a roughly 165-million-parameter language model, through a full training pipeline for about $4,200 in compute.","retrieval_nugget":"Empero says it trained and released Quark, a roughly 165-million-parameter language model, through a full training pipeline for about $4,200 in compute. The four Apache-2.0 checkpoints and detailed account make this useful to teams assessing what a small-budget model experiment can test—and what its results do not establish.","published_at":"2026-11-16","updated_at":"2026-11-16","record_date":"2026-11-16","date_kind":"published_at","topics":["compute"],"entities":["Empero","Quark","lium.io","Transformers"],"related_patterns":["agent-economics-moves-to-completed-work"],"novelty_label":"new_fact","what_changed":"Empero released Quark, a 165M-parameter model trained end to end for about $4,200, with 128K context, an RL-only version, and four Apache-2.0 checkpoints.","prior_refs":[],"source_url":"https://empero.org/writing/quark","source_urls":["https://empero.org/writing/quark"],"schema_version":"newruntime-agent-readable-v0.2","stable_id":"post:quark-our-whole-training-stack-end-to-end-eba988","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/quark-our-whole-training-stack-end-to-end-eba988/","markdown":"https://newruntime.com/posts/quark-our-whole-training-stack-end-to-end-eba988.md","json":"https://newruntime.com/posts/quark-our-whole-training-stack-end-to-end-eba988.json"},"markdown_url":"https://newruntime.com/posts/quark-our-whole-training-stack-end-to-end-eba988.md","next_reads":[{"type":"topic","path":"/topics/compute/","url":"https://newruntime.com/topics/compute/","title":"Compute","reason":"Explore the compute topic hub.","media_type":"text/html","markdown_url":"https://newruntime.com/topics/compute.md","json_url":"https://newruntime.com/topics/compute.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"}]}
