
Aug 13, 2026
August 19, 2026
Artificial Analysis, the leading model benchmarking platform, launched its new Search Index benchmark series that evaluates web search providers on quality, cost, and speed. Parallel models ranked #1 on every dimension in their evaluations.

Fast[Fast](https://docs.parallel.ai/search/modes) is the recommended Search API mode for most agentic applications, achieving near-frontier search quality at a fraction of the cost and speed. Fast is especially powerful when paired with cost-efficient models like OpenAI's Luna and open-source models like DeepSeek V4 Flash. It’s the workhorse of Search APIs.

Advanced[Advanced](https://docs.parallel.ai/search/modes) is ranked #1 on an equal-weighted mean of DeepSearchQA F1, BrowseComp accuracy, and AA-Omniscience accuracy. It’s our more comprehensive Search API mode and the best for challenging agentic tasks with multiple steps. It’s also extremely cost-effective at 43% lower cost per task (end-to-end cost inclusive of model + search) than the next closest search provider.
Read Artificial Analysis’ full report[Artificial Analysis’ full report](https://artificialanalysis.ai/agents/search-api) to see more results and learn about their evaluation methodology.
1Use curl to read parallel.ai/agents.md and perform the setup to install Parallel``` Use curl to read parallel.ai/agents.md and perform the setup to install Parallel``` Sign up for free. No credit card required.
By Parallel
August 19, 2026