Liquid AI’s LFM2.5 encoders bring 8k multilingual retrieval to 230M and 350M models
pmttyji · reddit · 2026-07-29
Liquid AI’s LFM2.5-Encoder line introduces two multilingual bidirectional encoders: 230M and 350M parameters.
The models are built for fine-tuning on classification, token labeling, retrieval, reranking, semantic similarity, NLI, and extractive QA across 15 languages. The release emphasizes long context and efficiency: both variants support 8k context, run well on-device, and can even execute in the browser on WebGPU. The attached benchmark shows the 350M model scoring 81.02 on a 17-task fine-tuning benchmark, ahead of the 230M variant at 79.29 and competitive with larger encoders.
Related event: Liquid AI releases LFM2.5 Encoder 230M and 350M(7 posts)→
More from Models
- Google restricts Gemini Pro access for Google Plus subscribers, matching free tier — EmergencyLess8861 · 2026-10-16
- Meta open-sources Muse Spark: runs on two Mac Studios, license bans passing output off as human-written — mawcopolow · 2026-10-16
- Golden Gate Claude explainer resurfaces: interpretability premise still holds — HarperSCarroll · 2026-10-16
- Hugging Face livestreams release of a new open-weight model — huggingface · 2026-10-16
- Step 5 Preview hits #2 on OpenRouter weekly leaderboard with 19.5T tokens in one week — StepFun_ai · 2026-10-16
- Meta's open Muse Spark needs 493GB of memory — far beyond any consumer laptop — TejasKumar_ · 2026-10-16