Search APIs allow AI agents to retrieve information from the web, grounding their responses in up-to-date sources. They are used across a wide range of agentic applications, including deep research, coding, and general knowledge work, to improve factuality and overall outputs. Search APIs differ in several important ways, including the underlying web index they search, how search results are presented to the model, and the tradeoffs they make between cost, speed, and retrieval quality. We benchmark 11 Search API products across 7 providers to help developers make more informed decisions when selecting a search tool for their next AI agent.
For benchmark definitions, scoring, and data handling details, see the methodology page.
Latest data: Aug 17, 2026.
Stirrup agentHighlights
Search-enabled answer quality across public benchmarks.
Candidate model cost and Search API cost per benchmark task and per search query.
Model time and search time per benchmark task and per search query.
Sortable public Search API benchmark rows and per benchmark breakdown.
Provider | ||||||||
|---|---|---|---|---|---|---|---|---|
 Parallel Search (advanced) | 75 | 42 | 81 | 77 | 67 | $47.93 | $35.58 | 37.5s |
 Exa Search (auto) | 74 | 41 | 78 | 74 | 70 | $65.57 | $61.58 | 27.8s |
 Firecrawl Search | 73 | 40 | 74 | 74 | 73 | $30.48 | $44.94 | 56.9s |
| Parallel Search (basic) | 73 | 40 | 79 | 73 | 68 | $45.14 | $69.69 | 22.2s |
| Exa Search (fast) | 68 | 35 | 76 | 61 | 69 | $78.11 | $80.53 | 22.6s |
 You.com Search | 68 | 35 | 63 | 74 | 66 | $68.93 | $58.28 | 35.8s |
| Parallel Search (turbo) | 67 | 34 | 70 | 75 | 56 | $13.64 | $46.57 | 20.7s |
 Keenable Search (pro) | 67 | 34 | 70 | 66 | 65 | $23.69 | $73.20 | 24.8s |
| Keenable Search (realtime) | 67 | 34 | 70 | 64 | 66 | $26.87 | $63.67 | 16.8s |
 Tavily Search (basic) | 66 | 33 | 74 | 59 | 64 | $126 | $66.58 | 38.6s |
 Brave Search | 65 | 32 | 62 | 67 | 65 | $71.61 | $74.52 | 28.9s |
Model only | 33 | baseline | 45 | 17 | 38 | — | $2.92 | 15.9s |
Representative example tasks for each benchmark behind the Artificial Analysis Search Index.
Broad research questions that need many searches. Answers are lists of items. An LLM grader scores each answer with an F1 score over the answer items. The full eval split has 900 tasks.
Representative example
I've completed the Desert Treasure quest on Old School Runescape, can you list all the names of the spells I can use to teleport into the wilderness that require more than four runes?
Answer: Dareeyak Teleport, Ghorrock Teleport
Hard-to-find facts that need multi-hop browsing. We use a hard 200-sample subset from the full evaluation pool. The grader checks for an exact answer.
Representative example
I'm looking for the name and location of a structure which fulfills the following criteria: 1. Located in Eastern Australia. 2. Can be visited on foot. 3. Was re-built in 2016. 4. Is longer than 50m. 5. Can be seen from another similar structure. 6. Hosts a yearly dinner. 7. Was originally built for another use case, but has not served that use case for a number of years.
Answer: Shorncliffe Pier, Brisbane, Australia
A private subset of 600 factual questions, balanced across 6 domains. The grader scores accuracy. The score shows how much search adds to the model's internal knowledge.
Representative example
In Karen Russell's short story "St. Lucy's Home for Girls Raised by Wolves," from which locality was the three-piece jazz band hired for the Debutante Ball?
Answer: West Toowoomba
The examples are representative problem types, not samples from the benchmarks.
Parallel Search (advanced) leads the Artificial Analysis Search Index at 75 across 7 Search API providers benchmarked by Artificial Analysis.
Per task, Keenable Search (realtime) is the fastest, with an average time per task of 16.8s (model time plus search time). Per individual search query, Keenable Search (realtime) is the fastest, with an average time per search query of 0.34s.
Parallel Search (turbo) has the lowest measured search cost at $13.64 per 1,000 benchmark tasks.
Search adds the most measured quality for Parallel Search (advanced), lifting the Artificial Analysis Search Index by 42 versus the same model with no search (its model-only baseline).
Search API answer quality is aggregated from public benchmarks including AA-Omniscience, BrowseComp, and DeepSearchQA. Each measures a model's ability to find, extract, or verify information using search.
The best provider depends on your priorities. Use the quality charts to compare answer accuracy, the cost charts to balance quality against search and model cost, and the latency charts for real-time use cases. The tradeoff scatter plots highlight the providers on the quality-cost and quality-latency frontiers. See the full methodology