tech
January 28, 2026
Exa AI Research Blog
Discover the latest in AI research and semantic search technology on the Exa blog. Learn how our neural network search engine provides high-quality web data for AI applications
TL;DR
- Exa 2.0 offers the fastest (sub-350ms) and highest quality search APIs.
- Exa Fast achieves sub-350ms P50 latency, 30% faster than competitors.
- Exa Auto is the default search type, optimized for both latency and quality.
- Exa Deep agentically searches and processes to retrieve the highest quality information (3.5s P50 latency).
- The updates are supported by a larger web index, a new embedding model, and vector database optimizations.
- Exa aims to provide near-instant search for AI grounding.
- Internal benchmarks and external benchmarks (like SimpleQA and Frames) were used for evaluation.
- The company is hiring engineers and researchers to build a superior search engine.