Kakao Kanana-2-embedding: open Korean-first embedding models for agent memory
Kakao open-sourced two embedding models on October 29, a 3B model and a 1.3B distilled version, with adjustable vector size for retrieval, aimed at finding past conversation context from vague queries like 'the quiet restaurant we talked about before'.
Why it matters
Embeddings built for personalized agent memory rather than generic search: they retrieve semantically similar past conversations, and Kakao claims they beat similarly sized open models in Korean while matching them in English.
What you could build with it
A team building a Korean-language support assistant could swap these embeddings into its retrieval layer and measure recall on real past tickets phrased vaguely by customers, comparing hit rate against its current embedding model before switching production search.
Does it hold up?
Too early to judge: released October 29 with vendor comparative-evaluation claims only; the models are on Hugging Face under an open license, but no independent retrieval benchmark or deployment report exists yet.
Built with Kakao Kanana-2-embedding
- Kakao open-sources Korean-focused embedding models to power AI searcharticle · Release-day report on the 3B and distilled 1.3B models, adjustable vector size, Kanana open license terms, and the Korean retrieval evaluation claims.
- mykor/kanana-2-3b-embedding-GGUFgithub · Community GGUF quant of the 3B embedding model for local retrieval pipelines.
Learn more
First spotted on hf: source.
More AI models
Tencent Hunyuan Hy3: open 295B MoE with 21B active under Apache 2.0Tencent open-sourced Hy3, a 295B-parameter mixture-of-experts model with only 21B parameters active per token, under…models · JEV 0.73NVIDIA Isaac Lab 3.0 GA + Isaac OS open runtime + Agile ONE SN humanoidMajor Isaac platform expansion at GTC Berlin Oct 21: Isaac Lab 3.0 is generally available as the open-source robot…models · JEV 0.73Kimi K3.1 Pro: Moonshot's 2.8T open-weight flagship officially launchesMoonshot AI officially launched Kimi K3.1 Pro on Oct 23, a mixture-of-experts model with 2.8 trillion total / 104…models · JEV 0.72NVIDIA Nemotron 3.5 Retriever collection: Embed 8B/1B, Rerank 8B, Parse 631MNVIDIA's Oct 21 collection drop for agentic retrieval: Nemotron-3.5-Embed in 8B BF16, 1B BF16 and 1B NVFP4…models · JEV 0.72NVIDIA Cosmos 3.5 world models: now openly availableNVIDIA's Cosmos 3.5 family of open world foundation models for robots, autonomous vehicles and vision agents is now…models · JEV 0.71NVIDIA RFD4-Proteina: open protein-design BioNeMo modelOpen protein-design model from NVIDIA and the Baker Lab at UW Medicine's Institute for Protein Design, announced at…models · JEV 0.7
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.