What matters in AI.

Subscribe

Models

LFM2.5-24B-A2B has an agent score above a model with more parameters

Its makers say that it has 48.09 on LiveClaw, against 38.19 for Qwen3.6-35B-A3B.

LFM2.5-24B-A2B vs larger models A checklist. The small model leads or matches on tool use and agents (5 of 8) and instruction following (3 of 3). Larger models lead on knowledge and coding. 16 benchmarks, 4 models. LFM2.5-24B-A2B vs larger models Tool use and agents 5 of 8 Leads or matches Instruction following 3 of 3 Leads Knowledge Larger models lead GPQA Diamond · HLE · AA-Omniscience Coding Larger models lead SWE-Bench Verified · SciCode 16 benchmarks, 4 models
The smallest model in a group of four leads on tool use and instruction following. Larger models still win on knowledge and coding.

Sources

  1. Run fast, private agents on-device with LFM2.5-24B-A2Bhuggingface.co
AI MATTER · NEWS · AI MATTER · NEWS ·13 OCT2026

Posted

Learn the terms in this story

More in Models

All Models news