Lateral Intelligence is a research lab advancing the next generation of language modelling.
We are working on enabling the next generation of language models with faster model inference. We're a small team based in San Francisco.
join@lateralintelligence.org
Research
All research →
Lateral Thinker
Our 130M param sequence latent refinement model with 72.1% on GSM8K.
Demo
Noise Your Prompt: Improved Conditional Generation in Diffusion Language Models
We offer a simple modification to training continuous diffusion language models that leads to improvements in both accuracy and generation diversity on combinatorial reasoning benchmarks.
Technical Report
Looped Flow Matching for Few‑Step Language Models
We improve continuous diffusion language models in the few-step regime by allowing them to correct previous integration errors, borrowing intuition from recursive reasoning models. On Sudoku hard this increases exact match accuracy from 38% to 72%.
Technical Report