Multilingual Retrieval Breakthrough: 307M Parameter Model Pushes Pareto Frontier
IgorCarron · x · 2026-07-31
Following the release of their SOTA English models DenseOn and LateOn, the team has extended their capabilities to the multilingual domain.
Experiments reveal that while mDenseOn is a strong model, the performance gap between late interaction (LI) and dense models remains massive on long contexts (like MLDR) and languages outside the training setup. This highlights the strong generalization capabilities of LI models. Furthermore, the new models push the Pareto frontier at 307M parameters across benchmarks including BEIR, MIRACL, MLDR, and MTEB 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