Liquid AI ships LFM2.5-24B-A2B: 24B MoE with 2B active params for local agents
JosephJacks_ · x · 2026-10-16
Liquid AI released LFM2.5-24B-A2B, the largest member of its LFM2.5 family, positioned as an efficient local model for laptops and PCs running multi-step tasks inside popular agent harnesses.
Key specs:
- 24B total parameters, 2B active per token
- Context expanded to 131K tokens, 128K vocabulary
- 33T pretraining tokens; it is a reasoning model
- Easy to fine-tune; open base and post-trained weights on Hugging Face
The company claims it performs competitively with Qwen3.6-35B-A3B — a model with 50% more total and active parameters — on agentic and tool-use benchmarks.
More from Models
- Thousands Hit by ChatGPT Standard Voice Mode Regression: Wrong Voice, Interruptions — shaunralston · 2026-10-16
- Gemini Argon 4 Quota Exhausted, Reset Countdown Shows Nearly 7 Days — ssh4net · 2026-10-16
- Ex-Microsoft exec Mikhail Parakhin: early Argon tests show strong math, top-notch post-training — MParakhin · 2026-10-16
- Anthropic readies work-email discount path to annual Claude Pro, checkout spotted — ryanmerket · 2026-10-16
- Claude now leads 26% of Anthropic's model R&D, up from under 1% in February — altryne · 2026-10-16
- Model flatly refuses a user's 'only lie from now on' simulation test prompt — Tischler285 · 2026-10-16