# SerpApi publishes benchmark comparing markdown search results across four extraction APIs

> SerpApi ran three identical queries through SerpApi, Tavily, Exa, and Firecrawl and compared markdown output location, token cost, and downstream model answer quality.

- Canonical: https://extractfeed.io/story/serpapi-publishes-benchmark-comparing-markdown-search-result-c0599ad/
- JSON: https://extractfeed.io/api/v1/stories/serpapi-publishes-benchmark-comparing-markdown-search-result-c0599ad.json
- Beat: Extraction & Parsing · Evidence: Primary source · Type/significance: analysis/2 · First seen: 2026-10-02T01:02:42.522023+00:00 · Updated: 2026-10-02T01:02:42.522023+00:00 · Edition: 2026-11-05
- Framing: model-written (headline, standfirst, why it matters, tags); source facts deterministic

## Briefing
- SerpApi: SerpApi ran three identical queries through SerpApi, Tavily, Exa, and Firecrawl and compared markdown output location, token cost, and downstream model answer quality.

## Why it matters
As LLM agents increasingly rely on markdown-formatted web content for grounding, the cost and structure of that markdown directly affect both latency and inference expense. This benchmark gives practitioners a rare apples-to-apples look at how four competing APIs differ on dimensions that matter for production agent pipelines, not just raw retrieval speed.

## Sources
- Primary source · SerpApi · 2026-10-01 — [Comparing Markdown Search Results: SerpApi vs. Exa, Tavily, and Firecrawl](https://serpapi.com/blog/markdown-search-results-serpapi-vs-exa-tavily-firecrawl/)

## Watch next
Will any of the benchmarked vendors publish a reproducible test harness so others can verify the results on their own queries?

Topics: SerpApi, Exa, Tavily, Firecrawl, markdown-extraction, benchmark, llm-agents, token-cost, search-apis

---
extractfeed is an agent-readable changefeed for web scraping and data extraction. Index: https://extractfeed.io/agents.md