Liquid AI releases LFM2.5-24B: just 2B active params per token for local agents
helloiamleonie · x · 2026-10-15
Liquid AI released LFM2.5-24B-A2B, the largest member of the LFM2.5 family, built for local multi-step agents on laptops and PCs. It's live on Ollama (207K+ downloads in 2 hours, 15GB at Q4 quantization).
- 24B total parameters with 2B active per token; a reasoning model with cheap thinking tokens
- Context expanded to 131K tokens, 128K vocabulary, 33T pretraining tokens
- Reported competitive with Qwen3.6-35B-A3B (50% more parameters) on agentic/tool-use tasks: LiveClaw 48.09 vs 38.19
- Multilingual (17 languages incl. Chinese), easy to fine-tune; open base and post-trained weights on Hugging Face under LFM Open License v1.0
Related event: LiquidAI Releases LFM2.5-24B-A2B, an On-Device MoE Reasoning Model(13 posts)→
More from Infra
- patio11 on multi-agent dev: clearly the future, and we'll buy so much inference — IgorCarron · 2026-10-15
- Hugging Face Hub makes model tensor inspection up to 20x faster, DeepSeek-V3.2 drops from 9.2s to 0.45s — victormustar · 2026-10-15
- RTX 5060 Ti 16GB surges up to 144% in Japan, hitting $1,292 as RTX 5080 tops $2,400 — chemist_slime · 2026-10-15
- Fish Audio Drama 3 lands in vLLM-Omni: 128 concurrent streams on one H200 — vllm_project · 2026-10-15
- 1,174 Macs form Darkbloom inference network, growing 4.5x in 60 days at half the cost — gajesh · 2026-10-15
- ByteDance in advanced talks with MediaTek for custom AI data-center chips, Bloomberg reports — mark_k · 2026-10-15