Benchmarks

Name: TinyFish benchmarks
Creator: TinyFish

Measured against the best alternative, every time.

TinyFish benchmark results against the strongest alternative on each metric. Sortable and filterable. Select a row to see the full field, methodology, and source.
	Leader vs. Runner up		Source
BrowseComp agentic research Average Search + Fetch latency	TinyFish1.47svsExa1.59s	1.1× faster	BrowseComp	Expand BrowseComp agentic research
TinyFish Fetch Quality Eval Pages returned as usable context	TinyFish93%vsTavily80%	13.0 pts ahead	Methodology	Expand TinyFish Fetch Quality Eval
Online-Mind2Web Success rate	TinyFish89.9%vsGemini 2.5 Computer Use69.0%	20.9 pts ahead	Online-Mind2Web	Expand Online-Mind2Web
Search latency p50 Search latency	TinyFish556msvsExa811ms	1.5× faster	—	Expand Search latency
SimpleQA First result contains enough evidence to answer	TinyFish49.2%vsTavily45.6%	3.6 pts ahead	SimpleQA	Expand SimpleQA
WebVoyager Task pass rate	TinyFish91.1%vsBrowserUse88.3%	2.8 pts ahead	WebVoyager	Expand WebVoyager
SimpleQA, end-to-end Questions answerable after Search + Fetch	TinyFish86.8%tiedParallel86.8%	Tied	SimpleQA	Expand SimpleQA, end-to-end
BrowseComp agentic research Tasks completed without timeout or error	Firecrawl96/100vsTinyFish95/100	1 pt behind	BrowseComp	Expand BrowseComp agentic research

Last updated Jul 2026 · all axes start at zero · margins measured against the strongest comparator

Give agents cleaner contact
with the live web.

Fresh sources from Search, clean context from Fetch, completed tasks from Web Agent.

No credit card required