LFM2.5-24B-A2B runs reasoning agents on personal devices
Hugging Face: Run fast, private agents on-device with LFM2.5-24B-A2B
Running large models locally without cloud servers could reshape how companies handle sensitive data and reduce latency costs.
Get smart on it
LFM2.5-24B-A2B is a reasoning model designed to run agents directly on personal devices while maintaining privacy. The model has 24 billion total parameters but uses only 2 billion active parameters per token, which reduces computational costs and enables fast inference on standard hardware. It performs competitively with larger models on tool use and agentic tasks, scoring particularly high on benchmarks measuring instruction following and reasoning capabilities, though larger models retain advantages on knowledge-heavy and coding tasks. The model is available for immediate use across multiple inference frameworks and supports reasoning, multi-step task execution, and tool calling within popular agent systems.
Relevant courses
Want to go deeper than the news? Explore live, cohort-based AI courses taught by practitioners.
Browse AI courses on MavenRelated news
Gemini 4 Argon: Built to solve your hardest problems
The new model aims to handle complex reasoning tasks where current systems often struggle or fail completely.
Also reported by The Verge, Ars Technica and The Decoder
Meta releases its first frontier model, Muse Wonder
Meta's Superintelligence Lab released a pair of new state-of-the-art models on Monday, Muse Wonder and Muse Wonder Full, the company's first attempt at a frontier model.
Also reported by The Verge, Wired AI, The Decoder and SiliconANGLE
Higgsfield takes AI video from 1 minute to over 10
Higgsfield brings creators longer videos with GPT-6 Astra, precise visuals with GPT-Image-2.5, and a path to voice filmmaking with GPT-Live.
The maker of non-text AI model Jev valued at $7.5B just weeks after launch
What has users and large corporations so excited about Jev is TypeSafe’s claim that it works significantly faster and uses far fewer tokens than LLMs.
Alibaba Qwen Introduces QwenGyre: Elastic Reinforcement Learning RL for XLong-Horizon Agents
Alibaba’s Qwen team has introduced QwenGyre, an elastic reinforcement learning RL framework for extreme-long-horizon agent training. QwenGyre is an online reinforcement learning system from Alibaba that trains coding agents through unmodified, black-box harnesses. It targets rollouts that run for hours and process close to 1M tokens. The QwenGyre elastic RL framework was used on Qwen’s flagship Qwen 3.8 2.4T model. It rests on 2 core ideas. It moves idle GPUs between rollout and tra
Odyssey-3 is a new generative world model that you can try for free
Odyssey is making its world model Odyssey-3 available as a public research preview. The 14-billion-parameter model generates interactive environments from text prompts in real time and can reportedly control robots, drones, and even GTA V. Odyssey claims top scores on physics benchmarks, though the methodology behind those results has limits. The article Odyssey-3 is a new generative world model that you can try for free appeared first on The Decoder.