Artificial Analysis (@ArtificialAnlys)T1·Tools
68 pts
Artificial Analysis benchmarked Parallel Search API's fast and advanced modes across 1,700 search tasks in its Stirrup agent harness, with advanced scoring 83 to tie Keenable Search (pro) atop the board and fast scoring 82 at fourth. Both modes gained 8-9 points since August, with fast mode costing just $0.014 per task and both hitting 89 F1 on DeepSearchQA.