
ZeroEntropy provides retrieval infrastructure for AI, including reranking, embedding, and search tools for production retrieval systems.
zeroentropy.devCrawled Oct 15, 2026
The page discusses why production search uses two models and what happens when they are paired.
Latest posts from the ZeroEntropy team on semantic search, reranking, and document retrieval.
Deep dives, tutorials, and technical explorations from the ZeroEntropy team.
ZeroEntropy offers pricing for semantic search and reranking, with a free starting tier and options to scale.
The page compares BGE's bge-m3 with zembed-1 on aggregate metrics measured on the eval set and says zembed-1 leads.
A head-to-head evaluation of ZeroEntropy's zembed-1 against all three Harrier embedding models across 28 diverse datasets, using graded LLM relevance judgments.
The page discusses open-source and open-weight rerankers as alternatives to closed-source models.
Head-to-head, data-backed comparisons of ZeroEntropy's embedding and reranking models against Cohere, Voyage, OpenAI, Jina, BGE, Harrier, and Gemini. All numbers sourced from /evals/.
ZeroEntropy trains small, calibrated models, including rerankers, embedders, classifiers, judges, and faithfulness scorers, for production AI systems.
ZeroEntropy trains small, specialized AI models, including rerankers, embeddings, and custom-trained models for production AI systems.
ZeroEntropy's Privacy Policy describes how the company collects, uses, and safeguards personal information.
ZeroEntropy's Terms of Use govern access to and use of its website, APIs, and services.
The tutorial shows how to use ZeroEntropy’s zerank-2-reranker to improve retrieval quality in retrieval-augmented generation (RAG) pipelines. It describes zerank-2 as a 4-billion-parameter, Qwen3-based cross-encoder reranker for a high-precision retrieve-and-rerank workflow.
May 2026 · MarkTechPostNews article
GitHubRepository owner
ZeroEntropy launches zembed-1, a multilingual text embedding model it describes as the world’s most accurate. The company says the model outperforms other embedding models by up to 7% on recall benchmarks.
Mar 2026 · ZeroEntropyNews article
ai71 and ZeroEntropy announce a partnership focused on enterprise document search, using advanced AI retrieval capabilities.
Nov 2025 · AI71News article