DFlash 2: Keep Drafting ParallelVisit link →DFlash 2 is the successor to our widely deployed parallel drafter: close to 3× the speed of autoregressive decoding, with the same output. Drafters for Qwen3.8-27B and Meta's Muse Glimmer are out today.August 22, 2026ai performancePermalink: 2026/w34/dflash-2-keep-drafting-parallel Copy Related LinksI ran 120B MoE models on my M1 Max ai performanceDroid 1.0: Building a Frontier Harness | Factory ai performanceGitHub - lithos-ai/lithos-metal ai performanceBest value LLM ai performanceGitHub - peonist-ai/halogen-flash-server ai performance← Back to Week 34
2026/w34/dflash-2-keep-drafting-parallel