LightOn Releases Open-Source 307M Multilingual Retrieval Models
antoine_chaffin · x · 2026-07-30
LightOn has released mDenseOn and mLateOn, two fully open-source 307M-parameter multilingual retrieval models. These models are trained on a massive translate-train corpus of 2.8 billion query-document pairs, supporting 9 natural languages and code retrieval.
Evaluations show that mLateOn achieves the best results on benchmarks like BEIR, target-language MIRACL, and MLDR. Furthermore, mLateOn demonstrates significantly better generalization to languages and scripts unseen during retrieval training compared to mDenseOn (e.g., averaging 67.59 vs 57.42 on MIRACL). The team has open-sourced the models, datasets, and training code.
Related event: LightOn Open-Sources Multilingual Retrieval Models mDenseOn and mLateOn(18 posts)→
More from Models
- Rumor: GLM 5.5 and Kimi K3.1 dropping soon, possibly slowed by anti-distillation blocks — zephyr_z9 · 2026-10-16
- Thrive Holdings benchmarks AI on real tax and IT work across 75+ firms — BorisMPower · 2026-10-16
- A Little Encouragement Makes a Small Local Model Draw a Stunning Pelican SVG — Fancy-Snow7 · 2026-10-16
- Rapid-MLX 0.16.2 brings NVIDIA Nemotron 3.5 up to 120B to Apple Silicon Macs — kuchaev · 2026-10-16
- "End of Chinese OSS" claim pushed back: open models get license almost automatically — eliebakouch · 2026-10-16
- Opus 5.5 still leads a month later, but open-weight decision models are closing in — airesearch12 · 2026-10-16