Perplexity came dead last after testing agentic search tools 3,537 times

Perplexity's agentic search tool ranked last out of eight services in a benchmark of 3,537 tests, with Serper, Tavily, and Parallel Pro taking the top spots. The Agentic Search Index measured accuracy, cost per 1,000 successful answers, and speed, providing a comparison for developers building AI agents.

Find the right tools for your AI agents Every tool claims to be the best. See which tools actually are best for your agents, starting with agentic web search. Agentic Search Indexi Methodology /methodology Agentic Search Index v0.1 · Jul 2026 Highlights Accuracy, cost, and speed for every tool, measured the same way. Agentic Search Indexi The overall score: how often each tool finds the right answer. Costi USD per 1,000 successful answers Speedi How long your agents wait for each search. Let Your Agent Access This Data Every number on this page is live over x402: an agent pays 0.035 USDC and gets the data with no account and no API key. A standard subscription based API is coming soon. Get data with one prompt → /agent-access Which search tool is right for your agents? No tool wins everything: match each tool's strengths to the questions your agents actually ask. Correct Answers by Question Typei From deep research to finding current info, each tool has its strengths. Failure Attributioni Know how each tool tends to fail before your agent depends on it. The bottom line Rankings on the left, real costs on the right. Every tool's score, cost, and speed in one table. | | Service | Index CI | Cost / 1k | p50 / p95 | |---|---|---|---|---| | 1 | | Serper /tools/serper-dev Tavily /tools/tavily-com Parallel Pro /tools/parallel-ai Exa /tools/exa-ai Brave /tools/brave-com Linkup /tools/linkup-so Parallel Base /tools/parallel-ai Perplexity Sonar /tools/perplexity-ai Cost Breakdowni The fee per call is often less than the token cost of reading what comes back.