LiquidAI/LFM2-1.2B-Longevity

TEXT GENERATIONConcurrent Unit Cost:1Model Size:1.2BQuant:BF16Context Size:32kPublished:May 25, 2026License:lfm1.0Architecture:Transformer0.0K Featherless Exclusive Cold

LiquidAI/LFM2-1.2B-Longevity is a 1.2 billion parameter language model developed jointly by Insilico Medicine and Liquid AI, featuring a hybrid Liquid architecture with multiplicative gates and short convolutions. Fine-tuned from LiquidAI/LFM2-1.2B, it is specifically adapted for interpreting heterogeneous aging biology data, utilizing a 32,768 token context length. This model excels in research applications related to aging biology and multi-omics interpretation.

Loading preview...

LiquidAI/LFM2-1.2B-Longevity: A Specialized Model for Aging Biology

LiquidAI/LFM2-1.2B-Longevity (L-LFM2-1.2B) is a 1.2 billion parameter language model developed by Insilico Medicine and Liquid AI. It is part of the Longevity-LLM (L-LLM) family, designed for interpreting diverse aging biology data.

Key Capabilities and Features

  • Domain-Adapted: Specifically fine-tuned on a comprehensive L-LLM corpus covering aging biology, including multi-omics and clinical data.
  • Hybrid Liquid Architecture: Built upon the LiquidAI/LFM2-1.2B base model, featuring a unique architecture with multiplicative gates and short convolutions.
  • Extended Context Length: Supports a substantial context length of 32,768 tokens, enabling processing of longer biological sequences and complex data.
  • Training Methodology: Utilizes full-parameter supervised fine-tuning with ChatML formatting and a dynamic-thinking template for flexible response generation.

Intended Use Cases

  • Aging Biology Research: Ideal for research applications focused on understanding and interpreting biological data related to aging.
  • Omics Interpretation: Particularly strong in analyzing and interpreting multi-omics data within the aging domain.

Important Considerations

  • Outputs are model predictions and require experimental validation; they are not intended as clinical advice.
  • Performance is optimized for data modalities present in its specialized training corpus.

This model is released under the LFM Open License v1.0, with the underlying base model retaining its original license.

Finetunes
1 models