project
Speculative decoding algorithm using an independent 1.4B draft model proposing fixed-size candidate blocks; the only configuration that reached usable throughput here.
No evidence-backed relationships are recorded.
Liquid AI shipped a 279.5-million-parameter draft model for LFM2.5-VL-3B on September 24, reporting 2.30x to 3.13x faster decoding on an Apple M5 Max. Image encoding and prompt prefill are unchanged, and the company's tests did not cover quantized deployments.
Reality
Inferact measured 709 output tokens a second on 16 Ironwood chips against 452 on 16 GB200s, at low concurrency with speculative decoding. The engineering worth reading is the hand-written memory schedule underneath.
Reality
DFlash uses a diffusion model to draft three tokens at once and claims to beat EAGLE-3. On one consumer GPU running llama.cpp on a JavaScript coding task, it trailed the drafter Google ships with Gemma-4-12B-it.
Reality
The release attaches speculative decoding to weights it publishes under MIT, which lowers what a self-hosting threat costs to stand up, even though the performance claim behind it is still the vendor's own.
Reality
The update adds a path selector and a two-tap convolution rather than layers, recovering most of the accuracy that tripling the drafter bought at 15.2% latency, by the vendor's own numbers.
Reality
A week-long failure log on two RTX 3090s under WSL2 lands on one config at 170-210 tok/s. Everything before it died in dependency resolution, not in the math.
Reality