以下部分引用 AI 总结，现阶段 AI 仍然有幻觉。请以内容的原文为准。

数据新鲜度群聊 07-14 ✓卖方 07-14 ✓Wrap 断档15天 ✕News 07-15 ✓X 断档12天 ✕页面生成 07-15

LFM2.5-Embedding-350M

6 atoms · 跨 1 天 · 首见 2026-06-18 · 最近 2026-06-18

三色: 🟦 fact 6 · 🟥 take 0 · stance ▲0/▼0/◆0

来源: X 6

时态: fresh:6

标签: 好数字:4 · 好思考:2

2026-06-18 · 产品发布
Introducing LFM2.5-Embedding-350M and LFM2.5-ColBERT-350M: two multilingual retrieval models built for ultra-fast and accurate search across 11 languages.
tags: 好思考 · → daily
2026-06-18 · 基础模型架构
Both are built on LFM2.5-350M-Base with bidirectional patches, the first full-context encoders in the LFM family.
tags: 好思考 · → daily
2026-06-18 · 检索延迟指标
End-to-end retrieval latency as low as 1.5ms with our enterprise stack! 🚀
tags: 好数字 · → daily
value: qty=1.5ms · date=2026-06-18
2026-06-18 · 多语言性能表现
Consistently best-in-class multilingual and cross-lingual performance across Arabic, German, English, Spanish, French, Italian, Japanese, Korean, Norwegian, Portuguese, and Swedish.
tags: 好数字 · → daily
2026-06-18 · 查询嵌入延迟（CPU/边缘设备）
Both run via llama.cpp GGUFs (CPUs, Laptops, edge devices) <10ms end-to-end query embedding latency.
tags: 好数字 · → daily
value: qty=<10ms · date=2026-06-18
2026-06-18 · 查询嵌入延迟（GPU）
Both fly on GPUs with our custom runtime <2ms end-to-end query embedding latency.
tags: 好数字 · → daily
value: qty=<2ms · date=2026-06-18