Summary: The Openbenchmarks company funding board tested 17 providers against the same domains and separated freshness from historical enrichment. Firecrawl’s agent led both measures, while stored-database providers generally lagged on new rounds but remained cheaper and faster. The article says teams should choose by latency, cost, and whether recent-stage accuracy actually matters.
Funding rounds do not hold still. One surfaces on a company newsroom, then the press picks it up, then it lands in a regulatory filing, and only after all of that does it arrive in a structured database. Every hop adds lag — which is why a provider that never misses a round from two years ago can still be flatly wrong about one announced last Tuesday.
The Openbenchmarks company funding board was built to measure exactly that. In August 2026, per Firecrawl’s write-up of the results, the independent board put 17 providers over the same company domains and checked every returned stage against a reviewed ground truth.
Two boards, because one rank hides the tradeoff
The board cuts its score into two halves. Freshness covers rounds announced in the trailing 30 days, asking how quickly a provider indexes a new announcement. Enrichment covers older rounds, asking how completely a provider has backfilled history.
The two pull in opposite directions. A provider that indexes fast may carry a thin historical record; a deep database may lag badly on last week’s news. Fold them into one averaged number and it would describe neither job, which is why the board ranks them separately.
One caveat belongs in every vendor conversation: correctness is judged on the latest stage alone. Round amount, total raised, and round count are counted for presence, not accuracy. A high “funding fields returned” figure says the fields were filled in — not that the values inside them are right.
What the freshness board shows
On recent rounds, providers reading the live web sit on top. Firecrawl’s agent posted 100% latest-stage-correct (Spark 2), followed by Exa’s instant search at 98.0%, Exa’s agent at 97.0%, Firecrawl’s second model at 96.0%, and Parallel’s Task API at 95.0%. Crunchbase’s reviewed export kept pace at 95.1%.
Stored-database GTM providers trail far behind on this board: Apollo at 59.7%, People Data Labs at 13.3%, CompanyEnrich at 12.7%. The gap is ingestion lag rather than incompetence. A database has yet to absorb a round announced nine days ago; an agent reading the live web already has it.
Where the stored records earn their place
Enrichment tightens the spread, yet agents still come out ahead. Firecrawl led enrichment at 92.3%, with Parallel at 90.0%, Exa’s deep and agent modes near 88.6%, and Crunchbase at 85.8%. Fiber was the strongest true-API GTM provider there at 84.9%.
Cost and speed invert the order. The accuracy leaders are also the slowest and the priciest per call: Firecrawl’s agent takes around 102 seconds and an estimated $1.98 per company, while Exa’s agent sits near $10.00. Web search APIs answer in seconds for cents, and GTM database lookups in a few hundred milliseconds. Those cost figures are modeled from observed billing units at published rates, not taken from invoices.
How to choose without a single winner
Choose by the job rather than by the ranking:
- High-value lookups and monitoring, where being right matters more than shaving milliseconds: a long-running agent.
- Volume screening, where a round from last week rarely matters: a database read is cheap and fast.
- The middle ground: web search APIs, whose accuracy tracks how well a round was covered on the open web.
If you are already wiring retrieval into an agent, the reasoning matches choosing a web search API for agents: define the retrieval task first, then let the benchmark axes — freshness, enrichment, latency, cost — pick the vendor.
The limitation is that this board judges a single field. If your product depends on round amounts or investor names being correct, the supplied results do not say which provider wins that job.