DFlash
by Z Lab · Model · Language & code models
DFlash 2 keeps the top 16 candidates per draft position, picks the best path with a lightweight selector, and adds a small convolution, gaining one more accepted token per pass for about 1% extra latency
Evidence
- Nov 14With its official 8-draft-token setting on SGLang, DFlash 2 used about 73 GB VRAM on an H100
- Sep 7On MI300X/MI355X in vLLM, DFlash reached up to 2.87x throughput on Gemma 4 26B-A4B versus 2.11–2.27x for EAGLE-3
- Sep 7DFlash gave 1.77–2.06x on Qwen3.6-35B-A3B and up to 2.68x on Kimi-K2.5 on AMD GPUs; native MTP beat it on Qwen3.5
- Aug 30DFlash2 support is awaited in Unsloth and LM Studio; reportedly could lift Mac throughput to around 30 tok/s
- Aug 19DFlash 2 speculative decoding layer for Qwen3.8-27B claimed to run nearly 3x faster with identical output
Show all 9
- Aug 19On an A100 with SGLang, Qwen3.8-27B ran at 59.1 tok/s with DFlash 2 versus 28.9 tok/s baseline, just over 2x, on five prompts
- Aug 11Meta ships a 1.6 GB DFlash drafter for Muse Glimmer that guesses 16 tokens ahead; the big model verifies, output identical to running alone
- Aug 11With the DFlash drafter, Muse Glimmer GGUF went from 31.5 to 53.17 tokens/s on an A6000, about 1.7x faster
- Jul 27Lucebox implements DFlash: draft model proposes up to 16 tokens, target verifies them in one forward pass