{"type":"post","id":"eagle-4-0-1-2-falls-back-to-4d2aee","slug":"eagle-4-0-1-2-falls-back-to-4d2aee","title":"EAGLE-4 0.1.2 falls back to plain SGLang decoding at high load and posts TokenSpeed tokens/s","description":"EAGLE-4 0.1.2 changes how its SGLang integration handles speculative decoding: it can skip speculation for a batch when plain decoding is faster.","retrieval_nugget":"EAGLE-4 0.1.2 changes how its SGLang integration handles speculative decoding: it can skip speculation for a batch when plain decoding is faster. On a four-B300 test system, the high-load recipe stayed near parity with plain decoding through 512 concurrent requests, while low-load configurations retained measured speedups.","published_at":"2026-11-08","updated_at":"2026-11-08","record_date":"2026-11-08","date_kind":"published_at","topics":["inference"],"entities":["EAGLE-4","eagle-harness","SGLang","vLLM","TokenSpeed","Qwen3.8-27B"],"related_patterns":["harness-architecture-outlives-model-choice"],"novelty_label":"new_fact","what_changed":null,"prior_refs":[],"source_url":"https://github.com/eagle-hq/EAGLE-4/blob/main/docs/release-notes/0.1.2.md","source_urls":["https://github.com/eagle-hq/EAGLE-4/blob/main/docs/release-notes/0.1.2.md"],"schema_version":"newruntime-agent-readable-v0.2","stable_id":"post:eagle-4-0-1-2-falls-back-to-4d2aee","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/eagle-4-0-1-2-falls-back-to-4d2aee/","markdown":"https://newruntime.com/posts/eagle-4-0-1-2-falls-back-to-4d2aee.md","json":"https://newruntime.com/posts/eagle-4-0-1-2-falls-back-to-4d2aee.json"},"markdown_url":"https://newruntime.com/posts/eagle-4-0-1-2-falls-back-to-4d2aee.md","next_reads":[{"type":"topic","path":"/topics/inference/","url":"https://newruntime.com/topics/inference/","title":"Inference","reason":"Explore the inference topic hub.","media_type":"text/html","markdown_url":"https://newruntime.com/topics/inference.md","json_url":"https://newruntime.com/topics/inference.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"}]}