Liquid AI releases LFM2.5-24B-A2B model with 24B parameters

Liquid AI: Liquid AI develops efficient AI models optimized for local deployment and agentic workflows on consumer hardware. The company has released its largest model in the LFM2.5 family, emphasizing ease of fine-tuning and strong performance on multi-step reasoning and tool-use tasks. This release positions the company as a provider of practical, open-weight solutions for developers working with agent harnesses. Hugging Face: Hugging Face operates a central platform for hosting, sharing, and deploying open machine learning models and related resources. The new LFM2.5-24B-A2B model has been uploaded there, enabling developers and researchers to access, fine-tune, and integrate the weights directly. This availability supports community-driven experimentation and adoption of the release. LFM2.5-24B-A2B: LFM2.5-24B-A2B is an open-weight reasoning model from the LFM2.5 family, designed as an efficient local model for laptops and PCs. It supports expanded context handling and a large vocabulary, making it suitable for customization and integration into agentic applications. The model is released in both base and post-trained versions for broad accessibility. Open Access: Both base and post-trained model weights are publicly shared on Hugging Face to support customization. Model Release: Liquid AI has launched a new open-weight reasoning model optimized for local multi-step agent tasks. Performance Focus: The model is positioned as competitive on agentic and tool-use benchmarks against larger alternatives.