---
type: "post"
id: "liquid-ai-has-published-lfm2-5-24b-a2b-d82305"
slug: "liquid-ai-has-published-lfm2-5-24b-a2b-d82305"
title: "Liquid AI has published LFM2.5-24B-A2B-DSpark on Hugging Face"
description: "Liquid AI has published LFM2.5-24B-A2B-DSpark on Hugging Face, a speculative-decoding draft model for LFM2.5-24B-A2B."
retrieval_nugget: "Liquid AI has published LFM2.5-24B-A2B-DSpark on Hugging Face, a speculative-decoding draft model for LFM2.5-24B-A2B."
published_at: "2026-10-23"
updated_at: "2026-10-23"
record_date: "2026-10-23"
date_kind: "published_at"
topics: ["inference","models","open-models"]
entities: ["Liquid AI","LFM2.5-24B-A2B-DSpark","LFM2.5-24B-A2B","SGLang","NVIDIA B300"]
related_patterns: []
novelty_label: "new_fact"
what_changed: "Adds the LFM2.5-24B-A2B-DSpark draft model, reporting up to 3.1× throughput on NVIDIA B300 and detailed concurrency benchmarks."
prior_refs: ["memory:3f494bf8-1e3f-8712-a9ab-60c04758a2e4","memory:46b4be50-c2b1-80f0-a956-d331ce7bfb3e"]
source_url: "https://huggingface.co/LiquidAI/LFM2.5-24B-A2B-DSpark"
source_urls: ["https://huggingface.co/LiquidAI/LFM2.5-24B-A2B-DSpark"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "post:liquid-ai-has-published-lfm2-5-24b-a2b-d82305"
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/liquid-ai-has-published-lfm2-5-24b-a2b-d82305/","markdown":"https://newruntime.com/posts/liquid-ai-has-published-lfm2-5-24b-a2b-d82305.md","json":"https://newruntime.com/posts/liquid-ai-has-published-lfm2-5-24b-a2b-d82305.json"}
---

# Liquid AI has published LFM2.5-24B-A2B-DSpark on Hugging Face

## Retrieval answer

Liquid AI has published LFM2.5-24B-A2B-DSpark on Hugging Face, a speculative-decoding draft model for LFM2.5-24B-A2B.

Liquid AI has published LFM2.5-24B-A2B-DSpark on Hugging Face, a speculative-decoding draft model for LFM2.5-24B-A2B. The target model verifies every token proposed by the draft model.

The repository lists 327.7 million BF16 draft parameters, five full-attention layers, a 128,000-token vocabulary, a rank-256 Markov head, a confidence head, and block size 9. Its SGLang example uses LFM2.5-24B-A2B-FP8 as the target and requires a recent build with LFM2-MoE DSpark support.

Liquid AI reports measurements on one NVIDIA B300 with an FP8 target and BF16 KV cache: arithmetic-mean output throughput was 2.38x, 2.69x, and 1.71x versus target-only decoding at concurrency 1, 32, and 256 respectively. The source does not establish task-quality gains, deployment cost, or equivalent results on other hardware. It also notes that floating-point differences can change generated text, including with greedy sampling.

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- topic hub: [Inference](https://newruntime.com/topics/inference.md): Explore the inference topic hub.
- topic hub: [Models](https://newruntime.com/topics/models.md): Explore the models topic hub.
- topic hub: [Open Models](https://newruntime.com/topics/open-models.md): Explore the open-models topic hub.
- collection: [Field Notes](https://newruntime.com/posts.json): Open the latest editorial analysis.
- collection: [Source Ledger](https://newruntime.com/source-ledger.json): Inspect the public source evidence graph.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)