搭載LFM2.5-24B-A2B，在裝置端快速運行私有AI代理
Back to Articles Run fast, private agents on-device with LFM2.
5-24B-A2B Team Article Published October 13, 2026 Upvote 5 Raahul Vignesh Raahul26 Follow LiquidAI Song Duong sduong Follow LiquidAI Akhilesh Deepak Gotmare akhileshdeepakgotmare Follow LiquidAI Fernando Fernandes Neto fernandofernandes Follow LiquidAI Jiahui Wang WWWjiahui Follow LiquidAI Mayank Agarwal mayankagarwal Follow LiquidAI nathan ranchin nathanrchn Follow LiquidAI Sergei Tilga tilgasergey Follow LiquidAI Sinoué GAD GAD-cell Follow LiquidAI Tim Seyde tseyde Follow LiquidAI Zetian Li zetianli Follow LiquidAI changyi yang changyiyang Follow LiquidAI Simon Lee yungisimon Follow LiquidAI Leonie Monigatti iamleonie Follow LiquidAI Maxime Labonne mlabonne Follow LiquidAI Karthik Macherla karthikrmacherla Follow LiquidAI Alfred Eng alfredeng Follow LiquidAI Khoi Nguyen KTLK Follow LiquidAI Paul Pak paulpak58 Follow LiquidAI xx tugot17 Follow LiquidAI Tarek Dakhran tdakhran Follow LiquidAI Anthony Maltsev amaltsev Follow LiquidAI LFM2.
5-24B-A2B is a 24B-parameter reasoning model with just 2B active parameters, built for fast, private agents on-device.It thinks before it answers, calls tools, and works through multi-step tasks inside popular agent harnesses.
Because only 2B parameters are active per token, each reasoning token is cheap, so you can run agents continuously on your own hardware.Best-in-class agent: Competitive with Qwen3.6-35B-A3B, a model with ~50% more total and active parameters, on agentic and tool-use tasks, scoring 48.
09 on LiveClaw versus 38.19.Reasoning at MoE speed: Always thinks before it answers.With only 2B active parameters per token, each reasoning token is cheap.Efficient inference: 184 tok/s on an Apple M5 Max and up to 21K output tok/s on a single NVIDIA B300 GPU.
How we built a reasoning agent with 2B active parameters LFM2.5-24B-A2B builds on LFM2-24B-A2B and keeps its architecture: 40 layers, 64 experts per MoE block with top-4 routing, and the same hybrid of gated short convolutions and GQA.
The context window grows from 32K to 131K tokens, and an in-place tokenizer expansion doubles the vocabulary from 65K to 128K tokens for better non-Latin tokenization.Unlike its predecessor, it is a reasoning-only model that produces an explicit chain of thought before its final answer.
Pre-training was scaled up from ~17T to ~33T tokens.Mid-training then adds 2T tokens at 32K context, weighted toward reasoning, math, code, instruction following, and multilingual text, and 300B tokens at 131K context, weighted toward long-context data.
Tool calls and multi-turn trajectories appear throughout both phases.Post-training turns the base model into an agent in five stages: Supervised fine-tuning (SFT) Teacher specialization Multi-domain on-policy distillation (MOPD), non-agentic: distill the specialist teachers on non-agentic tasks.
MOPD, agentic: a second distillation stage focused on agentic tasks.Agentic reinforcement learning (Agentic RL): train the model inside agent harnesses, such as OpenClaw, Hermes Agent, Pi, and OpenCode.Benchmark results We evaluated LFM2.5-24B-A2B against Qwen3.
6-35B-A3B, Gemma-4-26B-A4B-it, and Nemotron-3.5-Lightning on 16 benchmarks covering tool use and agentic workflows, knowledge, instruction following, math, and coding.It is the smallest model in the group, yet it is competitive with MoEs with ~50% more total and active parameters.Benchmark LFM2.
5-24B-A2B Qwen3.6-35B-A3B Gemma-4-26B-A4B-it Nemotron-3.5-Lightning (30B-A3B) AutomationBench 27.69 8.72 5.46 2.61 BFCLv4 65.01 65.00 57.09 53.78 ToolSandbox 80.71 75.02 72.09 75.13 LiveClaw 48.09 38.19 41.25 37.72 PinchBench 83.18 89.93 77.12 84.20 τ³-Bench Banking 8.04 7.63 4.54 7.
01 WebArena Verified 26.75 37.62 25.62 28.69 GPQA Diamond 74.42 83.69 74.90 72.42 AA-Omniscience -24.90 -24.35 -59.39 -3.58 HLE 7.40 7.69 8.57 5.65 IFEval 96.75 93.06 92.00 94.19 IFBench 74.69 64.21 47.44 71.67 Multi-IF 84.38 81.53 82.19 82.34 MATH-500 96.04 96.73 95.66 96.
40 SWE-Bench Verified (OpenCode) 51.00 66.20 25.80 53.40 SciCode 32.71 37.35 35.42 27.78 *All the models were evaluated with vLLM, using each benchmark's recommended generation parameters.For your app, the strengths are tool use and instruction following.LFM2.
5-24B-A2B leads or matches the competition on five of the eight tool-use and agentic benchmarks (AutomationBench, BFCLv4, ToolSandbox, LiveClaw, and τ³-Bench Banking), and on AutomationBench it scores more than three times the next-best model.
It also leads on all three instruction-following benchmarks.Larger models keep an edge on knowledge-heavy benchmarks like GPQA Diamond, HLE, and AA-Omniscience, and on coding benchmarks like SWE-Bench Verified (OpenCode) and SciCode, so pair LFM2.
5-24B-A2B with retrieval or web search for knowledge-heavy workloads.These results make LFM2.5-24B-A2B a strong foundation for agents that need to reason, call tools, and run reliably on your device.Inference speed on CPU and GPU LFM2.
5-24B-A2B ships with day-one support across the inference ecosystem, including llama.cpp, MLX, vLLM, SGLang, and ONNX.Edge inference.LFM2.5-24B-A2B is the fastest of the compared MoEs, decoding 184 tokens/s on an M5 Max and 114 tokens/s on a Ryzen AI Max+ 395.That is up to 1.
7x faster than the next-best model, while using the least memory.GPU inference.LFM2.5-24B-A2B is the fastest model in its size class, reaching over 21K output tokens per second at high concurrency on a single B300.How to use LFM2.5-24B-A2B Reach for LFM2.
5-24B-A2B when you need agents that reason, call tools, and run reliably on your device.Install the latest version of transformers (compatible with transformers>=5.0.
0): pip install -U transformers Then load and run the model: from transformers import AutoModelForCausalLM, AutoTokenizer, TextStreamer modelid = "LiquidAI/LFM2.5-24B-A2B" model = AutoModelForCausalLM.
frompretrained( modelid, devicemap="auto", dtype="bfloat16", # attnimplementation="flashattention2" <- uncomment on compatible GPU ) tokenizer = AutoTokenizer.frompretrained(modelid) streamer = TextStreamer(tokenizer, skipprompt=True, skipspecialtokens=True) prompt = "What is C.elegans?
" inputids = tokenizer.applychattemplate( [{"role": "user", "content": prompt}], addgenerationprompt=True, returntensors="pt", tokenize=True, )["inputids"].to(model.device) output = model.generate( inputids, dosample=True, temperature=0.1, topk=50, repetitionpenalty=1.
05, maxnewtokens=512, streamer=streamer, ) Use it in your agent harness Serve LFM2.5-24B-A2B locally behind any OpenAI-compatible backend, then point your agent harness at it.For example, with Hermes Agent: hermes config set model.provider custom hermes config set model.
baseurl http://localhost:8080/v1 hermes config set model.default LFM2.5-24B-A2B hermes config set model.contextlength 131072 hermes config set model.apimode chatcompletions hermes config set agent.tooluse_enforcement true The port depends on your serving backend: llama.
cpp and MLX use 8080, vLLM uses 8000, SGLang uses 30000, and LM Studio uses 1234.See the model card for OpenClaw and Pi configs, and our guide for the full setup.Check out our release blog for video examples.Get Started with LFM2.5-24B-A2B Both LFM2.5-24B-A2B and LFM2.
5-24B-A2B-Base are open-weight and available on Hugging Face today: Download: LFM2.5-24B-A2B-Base and LFM2.5-24B-A2B on Hugging Face.Use in your harness: follow our guide on how to run a local agent, like OpenClaw, Hermes Agent, and Pi.Fine-tune: adapt LFM2.
5-24B-A2B to your task with our fine-tuning tutorials.We can't wait to see what you build.Citation If you use this work, please cite the release blog: @article{liquidAI202624BA2B, author = {Liquid AI}, title = {LFM2.
5-24B-A2B: Built for Fast On-Device Agents}, journal = {Liquid AI Blog}, year = {2026}, note = {www.liquid.
ai/blog/lfm2-5-24b-a2b}, } Models mentioned in this article 3 More from this author Multimodal open d1 decision models for the edge 38 October 7, 2026 Accelerating vision-language models with LFM2.
5-VL-DSpark 51 September 24, 2026 Community EditPreview Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.Tap or paste here to upload images Comment · Sign up or log in to comment Upvote 5 Models mentioned in this article 3
Related
相關文章
TinyFish Makes Search Free: What It Really Costs an AI Agent to Read the Web
An AI agent that reads the web pays 2 kinds of cost.The first is the retrieval charge: what a search or fetch API bills per request.
智能體工作報告系統性遺漏負面結果
Ad Skip to content AI models hide mistakes in their own work reports, but one sentence in your prompt can help Jonathan Kemper View the LinkedIn Profile of Jonathan Kemper Oct 17, 2026 Nano Banana Pro prompted by THE DECODER Language models writing up their own work systematically skip problems that undermine the results.
Perplexity Open-Sources SPACE, a Firecracker MicroVM Sandbox for AI Agents
Perplexity has open-sourced SPACE, the sandbox platform behind every task in Perplexity Computer. The Perplexity SPACE sandbox gives each agent task its own Firecracker microVM. Credentials and outbound traffic are controlled from outside that VM.
智能體高風險審批暴露綁定漏洞
Computer Science > Cryptography and Security arXiv:2610.21063 (cs) [Submitted on 15 Oct 2026] Title:Approve What, Exactly? Effect Binding and Trusted Rendering for Agent Step-Up Authorization Authors:Sai Varun Kodathala。
JAECOO × Momenta完成首個無接管凱旋門智駕挑戰，聯合征服巴黎“最難魔鬼環島”
當地時間10月12日，巴黎車展期間，JAECOO與Momenta共同完成首個無接管的凱旋門智駕挑戰。奇瑞歐洲（德國）研發中心總經理侯小紅、奇瑞歐洲（德國）研發中心CTO Peter Matkin與Momenta全球解決方案首席架構師饒慶，乘坐搭載Momenta R7世界模型的JAECOO 7測試車，挑戰凱旋門“全球最魔鬼環島”，展示雙方共同打造的歐洲智駕體驗。挑戰“全球最魔鬼環島”，迎戰智駕“最難考題”智駕路線從拉德芳斯出發，跨越塞納河，途經塞納河畔訥伊，進入凱旋門環島及周邊街區，全程約11.38公里。
AI 代理網購遭亞馬遜封殺！馬斯克被逼急：大不了自己建個電商帝國去競爭
馬斯克被逼急：大不了自己建個電商帝國去競爭發布於AI新閒資訊時間 :Oct 16, 2026閱讀 :1分鐘人工智能智能體（AI Agent）正試圖顛覆傳統的網購模式，但各大傳統電商平臺顯然並不打算輕易敞開大門。近日，有用戶在社交平臺發帖反饋，稱亞馬遜正在通過技術手段阻止埃隆·馬斯克旗下的 AI 智能體 Grok Bot 在其平臺上替用戶完成下單操作。