Details
- Perplexity reports that its Search API now holds the top three positions on the Artificial Analysis Search Index.
- The benchmark evaluates search APIs on quality, cost per task, and related efficiency metrics for agent-style workloads.
- Perplexity’s medium Search API setting is highlighted as scoring five points higher than the previous leading provider.
- At the medium setting, the Search API is said to extend the quality-cost Pareto frontier at roughly $0.091 per task.
- The result suggests that Perplexity’s medium configuration offers improved answer quality at a competitive cost compared with earlier top-ranked services.
- The low and high context variants of the same endpoint also participate in the benchmark, filling out the top three spots and underscoring consistency across configurations.
- These scores are based on Artificial Analysis’s standardized harness, which swaps only the search provider while holding the underlying model and agent logic constant.
Impact
Perplexity’s sweep of the Artificial Analysis Search Index strengthens its position in the search API market and pressures rivals that previously led this benchmarked landscape. By pushing the quality-cost Pareto frontier at around $0.091 per task, Perplexity makes high-quality, agent-oriented web search more economically attractive, which could accelerate adoption of retrieval-heavy AI agents and force competing search providers to adjust pricing, performance, or both to stay credible in technical evaluations.
