Modellfm2:24b
LFM2-24B-A2B
The previous-generation 24B total-parameter mixture-of-experts model from Liquid AI. It uses a hybrid convolution and GQA architecture with 40 layers and 64 experts under top-4 routing, supports a 32K context, and is the predecessor of LFM2.5-24B-A2B. It can be run on Ollama.
Timeline (last 90 days)
Oct 14
Compared with the previous generation LFM2-24B-A2B, LFM2.5-24B-A2B's architecture is unchanged (40 layers, 64 experts, top-4, convolution/GQA hybrid), with the main upgrades in training scale, long context, and reasoning capability: context increased from 32K to 131K, vocabulary from 65K to 128K, pre-training tokens from 17T to 33T, plus a newly added always-on chain of thought and 5-stage post-training (ending with agentic RL).
Unverified50%