Best AI News today

LFM2.5-DSpark — AI News Today

Science

18 stories about LFM2.5-DSpark

1 from your feeds4 from search

LFM2.5-DSpark — Coverage Momentum

2
This month
—
vs last month
2
All-time
Aug 2026
Peak month
0
0
0
0
0
2
MarAprMayJunJulAug

First tracked Aug 2026. Most-citing sources: Hugging Face (1), Marktechpost (1). Data aggregated by Best AI News Today from 30+ sources.

18 stories
Liquid AI Releases LFM2.5-DSpark Draft Models That Deliver Up to 3.18x Faster Decoding Without Changing Model Outputs
Top StoryMedia

Liquid AI Releases LFM2.5-DSpark Draft Models That Deliver Up to 3.18x Faster Decoding Without Changing Model Outputs

Liquid AI has cracked a problem that has long frustrated AI developers: how to make large language models run dramatically faster without forcing companies to retrain their systems or accept degraded outputs. The startup's new LFM2.5-DSpark models achieve up to 3.18 times faster decoding speeds while producing identical results to their standard counterparts, a feat that could reshape the economics of deploying AI in cost-sensitive industries like banking and e-commerce. If the claims hold at scale, the breakthrough suggests a pathway to making cutting-edge AI affordable for thousands of smaller companies that have until now been locked out of the market.

MarkTechPost faviconMarkTechPost·

5 sources

TUN - The University Network faviconTUN - The University NetworkUnite.AI faviconUnite.AIfinance.biggo.com faviconfinance.biggo.com
AIが最大3.2倍高速に、LFM2.5用ドラフトモデル無償公開
Media

AIが最大3.2倍高速に、LFM2.5用ドラフトモデル無償公開

ノートPCなどの上で動くAIエージェント特化モデル「LFM2.5」がさらなる高速化を果たした。Liquid AIは8月20日、LFM2.5ファミリーの推論速度を大幅に引き上げる「DSparkドラフトモデル」をHugging Faceで公開した。LFM(Liquid Foundation ...

PC Watch on MSN faviconPC Watch on MSN·
小型モデル「LFM2.5」を2倍以上高速化する投機的デコーディング技術DSpark適用版が登場
Media

小型モデル「LFM2.5」を2倍以上高速化する投機的デコーディング技術DSpark適用版が登場

AI企業のLiquid AIが小型モデルのLFM2.5シリーズに高速化技術の「DSpark」を適用したことを発表しました。DSpark適用版のモデルは性能を維持したまま2倍以上の高速化に成功しています。

GIGAZINE on MSN faviconGIGAZINE on MSN·
Media

No cloud, no GPUs, no problem: Liquid AI's new model LFM2.5-2.6B brings powerful AI agents to devices as small as a Raspberry Pi

Credit: VentureBeat made with OpenAI ChatGPT-Images-2.0 Earlier this week, the AI startup Liquid, formed in 2023 by former MIT computer scientists, debuted LFM2.5-2.6B, a new open-weight language ...

VentureBeat faviconVentureBeat·
Media

A compact AI model with Japanese language support that also works on smartphones, the 'LFM2.5-8B-A1B,' has been released.

GIGAZINE faviconGIGAZINE·
Media

Liquid AI Released LFM2.5-350M: A Compact 350M Parameter Model Trained on 28T Tokens with Scaled Reinforcement Learning

MarkTechPost faviconMarkTechPost·
Media

Introducing the on-device-focused open model 'LFM2.5' series, including the small AI model 'LFM2.5-1.2B-JP' specialized for Japanese

GIGAZINE faviconGIGAZINE·
Media

AI Industry's GitHub Seeks Sale, Hugging Face Valuation May Exceed $13 Billion

AIBase faviconAIBase·
Research

Up to 3.2x Faster Inference with LFM2.5-DSpark

Hugging Face faviconHugging Face·
Media

Liquid AI Releases LFM2.5-2.6B: An On-Device Agentic Model With 128K Context, Tool Calling, And Open Weights

MarkTechPost faviconMarkTechPost·
Media

Liquid AI Releases LFM2.5-VL-450M: a 450M-Parameter Vision-Language Model with Bounding Box Prediction, Multilingual Support, and Sub-250ms Edge Inference

MarkTechPost faviconMarkTechPost·
Media

Liquid AI Releases LFM2.5-1.2B-Thinking: a 1.2B Parameter Reasoning Model That Fits Under 1 GB On-Device

MarkTechPost faviconMarkTechPost·
Media

Microsoft's OmniParser Open Source Project Rises to the Top of HuggingFace's Most Popular Models

AIBase faviconAIBase·
Media

Anthropic Brings Claude Mythos 5 to Claude Security: Enterprise Teams Get Frontier Vulnerability Scanning Without Direct Model Access

MarkTechPost faviconMarkTechPost·
Media

A QAD (Quick Adapter) version of the 'LFM2.5,' a compact AI model that runs on smartphones, has been released, featuring reduced memory usage while minimizing performance degradation.

GIGAZINE faviconGIGAZINE·
Media

Liquid AI's smallest model yet LFM2.5-230M beats models 4X its size at data extraction, can run 'anywhere'

Liquid AI, founded by former MIT computer scientists, today released its smallest AI language model yet, LFM2.5-230M, and enterprises would do well to consider it for their uses in data extraction...

VentureBeat faviconVentureBeat·
RelatedMIT
Media

ACE-Step 1.5 XL, a locally executable music generation AI, has been released, capable of generating songs with Japanese vocals.

GIGAZINE faviconGIGAZINE·
RelatedAI & Art
Media

1 to 8! Alibaba Qwen's Download Volume Leads by a Large Margin, Beating the Total of Global Giants Like Meta and OpenAI in a Single Month

AIBase faviconAIBase·

Frequently Asked Questions

What is LFM2.5-DSpark?

LFM2.5-DSpark is an AI science model. Best AI News Today aggregates the latest news, benchmarks, and developments from 30+ AI sources.

What is the latest news about LFM2.5-DSpark?

As of today, there are 18 recent stories about LFM2.5-DSpark. Recent headlines include: AIが最大3.2倍高速に、LFM2.5用ドラフトモデル無償公開; AI Industry's GitHub Seeks Sale, Hugging Face Valuation May Exceed $13 Billion; Anthropic Brings Claude Mythos 5 to Claude Security: Enterprise Teams Get Frontier Vulnerability Scanning Without Direct Model Access. Updated every 15 minutes.

What other Science AI models are there?

LFM2.5-DSpark is part of the Science model category. Browse all comparable AI science models with live news mentions and benchmarks.