MinishLab ships Blink static embedding models: 8M and 64M params, fast CPU retrieval
tomaarsen · x · 2026-11-18
MinishLab released its Blink static embedding models: an 8M-param English model (blink-8M) and a 64M-param multilingual model covering 100+ languages, both outputting 256-dim vectors under an MIT license.
The models simply look up token vectors, average and normalize them — no transformer forward pass — so they run fast on CPU with 31MB/256MB of float32 weights, and load directly via Sentence Transformers. The attached demo shows blink-8M on CPU correctly matching a password-reset query to the right help article.
Related event: MinishLab Releases Blink Static Embeddings for Fast CPU Retrieval(5 posts)→
More from Models
- Raschka: benchmarking new LLMs is like using IQ tests for hiring — rasbt · 2026-11-18
- Open-source Karpo-Decision-31B tops 300-model Decision Index with zero-token, one-forward-pass answers — rohanpaul_ai · 2026-11-18
- GPT-6.2 backlash is real user experience, not astroturf: dev says Astra has quirks but Sol is a solid daily driver — pvncher · 2026-11-18
- LLM protein-folding test on fresh PDB entry lands near-random structures — MoAlQuraishi · 2026-11-18
- Every's founder pushes back: we re-tested the model to the wire, OpenAI gave us no heads-up — kieranklaassen · 2026-11-18
- HF trending: Thinking Machines' Inkling-2-Small, MiniMax H3-Image and Qwen3.8-27B surge together — victormustar · 2026-11-18