Exa Launches Agent Ultra: A Subagent Swarm Deep Research API Built for Exhaustive List Building
Exa's Agent Ultra coordinates thousands of sources through subagent swarms — and claims to outperform Opus 5.5 and GPT-6 Astra on list-building benchmarks.
Why it matters
A new agent API tier targeting exhaustive entity research and list building. Exa benchmarks the swarm approach against frontier models on recall-heavy tasks; practitioners should evaluate whether the subagent coordination model improves over single-agent retrievals for their research and enrichment workflows.
The key facts
5 to knowExa Agent Ultra uses subagent swarms across thousands of sources
81.4% soft recall on WANDR benchmark
Claims to beat Opus 5.5, GPT-6 Astra, and Perplexity Agent on 4 benchmarks
Focused on list building and entity enrichment
Product positioning: 'highest effort mode' of Exa Agent API
Go to the source
MarkTechPostmarktechpost.com
Publisher excerpt: Exa has released Agent Ultra, the highest effort mode of its Exa Agent API. It coordinates subagents across thousands of sources for list building and entity enrichment. Exa reports it beats Opus 5.5, GPT-6 Astra, and Perplexity Agent on 4 benchmarks, including 81.4% soft recall on WANDR.