DFlash 2 squeezes 16-25% more tokens per pass from speculative decoding, for ~1% latency
Inco AI's DFlash 2 pushes parallel drafting further: 16-25% more output per verification pass at ~1% added cycle latency, with provably unchanged output. Already live in SGLang, vLLM, TensorRT-LLM, and llama.cpp, it serves single-batch LLM inference at 2.7-3.4x autoregressive throughput.