Exa Launches Deep Max Search API, Claims 20x Faster Than Competitors at Equal Accuracy

CoinDesk reports that on April 21 (UTC+8), according to monitoring by Beating, Exa has launched Deep Max, a high-end search interface designed for LLM agents. Each query triggers dozens of concurrent Exa searches, each approaching the same question from a different angle, before consolidating the results into a single answer. Agentic search refers to LLMs autonomously planning multi-step searches and synthesizing responses—a primary product format adopted over the past year by search API companies such as Perplexity, You.com, and Parallel. Three benchmarks were conducted internally by Exa, comparing Deep Max against competing products on identical queries: - Deep Search QA: Deep Max achieved 90% accuracy in 64 seconds; You.com Frontier scored 84% in 5,908 seconds; Parallel Ultra 8x scored 82% in 1,703 seconds. - FRAMES: Deep Max achieved 94% accuracy in 11 seconds; Parallel Ultra scored 88% in 1,457 seconds. - HLE-Search: Deep Max achieved 80% accuracy in 25 seconds—matching GPT-5.4’s accuracy while halving the response time. Other native search tools compared included Gemini 3.1 Pro and Claude Opus 4.7. Exa claims Deep Max is “up to 20 times faster than the closest competitor.” This speed is attributed to parallel tool invocation, compressed page content, and Exa’s proprietary search stack, which returns results within one second per call. Deep Max is currently available only to customers in contact with sales; the official release date and pricing have not been disclosed.