← all releases
LightOn·Feb 19, 2026·v1.0·open source
ColBERT-Zero
ColBERT-Zero is a late interaction model trained from scratch using PyLate, designed for multi-vector retrieval tasks. Developed in collaboration with École Polytechnique Fédérale de Lausanne (EPFL) and the Swiss AI Initiative, it achieves state-of-the-art results on the BEIR benchmark with under 150 million parameters, utilizing only public data.
ML
overview
ColBERT-Zero is a late interaction model trained from scratch using PyLate, designed for multi-vector retrieval tasks. Developed in collaboration with École Polytechnique Fédérale de Lausanne (EPFL) and the Swiss AI Initiative, it achieves state-of-the-art results on the BEIR benchmark with under 150 million parameters, utilizing only public data.
key features
- 01Late interaction model architecture
- 02Trained from scratch with PyLate
- 03State-of-the-art performance on BEIR benchmark
- 04Utilizes only public data for training
- 05Collaboration with EPFL and Swiss AI Initiative
related productsbrowse all →
SkyRL txopen sourceSkyRL tx is an open-source library that implements a backend for the Tinker API, enabling users to set up their own Tinker-like services on personal hardware. It supports end-to-end reinforcement learning (RL) and offers significantly faster sampling. The library is designed to be modular, allowing easy prototyping of new training algorithms, environments, and execution plans without compromising usability or speed.Step-Audio-R1open sourceStep-Audio-R1 is an advanced audio language model developed by StepFun AI, designed to enhance audio reasoning capabilities by grounding its reasoning in acoustic features. It introduces Modality-Grounded Reasoning Distillation (MGRD), an iterative training framework that shifts the model's reasoning from textual abstractions to acoustic properties, effectively addressing the 'inverted scaling' problem where performance degrades with longer reasoning. This model has demonstrated superior performance across various audio understanding and reasoning benchmarks, surpassing models like Gemini 2.5 Pro and achieving results comparable to Gemini 3 Pro.SINQopen sourceSINQ (Sinkhorn-Normalized Quantization) is a novel, fast, and high-quality quantization method designed to make any Large Language Model (LLM) smaller while preserving accuracy. It offers a plug-and-play, model-agnostic technique that delivers state-of-the-art performance for LLMs without sacrificing accuracy.