Skip to main content
We ran a first-party benchmark comparing selected SEC API operations against comparable sec-api.io operations. In this March 18, 2026 run, SEC API returned lower p50 latency and smaller response payloads for the measured entity-resolution, filing-search, section-extraction, and XBRL-conversion tasks.

​
Methodology

This is a first-party head-to-head benchmark run by our own canonical harness, not a third-party study. Our production workflow uses GitHub Actions ubuntu-latest. Within each workflow run, comparands execute sequentially on the same runner, which GitHub assigns from a variable pool.
  • Competitor measurements captured 2026-03-18; re-run before citing as current
  • Hosted APIs tested from the same GitHub Actions runner within a workflow run
  • Each operation run 5+ times with p50, p95, p99 latency reported
  • Payload size measured as response body bytes
  • Token estimate = ceil(payload_bytes / 4)
  • All tests use authenticated API keys on production endpoints
  • Reproduction scripts available in our benchmark methodology notes

​
Results

​
Latency Comparison

Overall: 18 wins, 0 losses, 2 ties across our benchmark suite.

​
Token Efficiency

​
How To Interpret Token Efficiency

When AI agents query SEC data, every byte of the API response consumes tokens. Smaller responses mean:
  • Lower cost — fewer tokens = lower API bills
  • Faster reasoning — less context for the model to process
  • More room — more data fits in the context window
Do not extrapolate these response-size measurements to other agent workflows. A company-briefing, filing-monitor, or RAG pipeline may have different call counts, retrieval steps, prompts, and output shapes. Re-run the benchmark or instrument your own workload before making cost or latency claims for a production workflow.

​
Pricing Comparison

For light agent usage (200 queries/month):
  • sec-api.io: $49/month (minimum tier)
  • SEC API PAYG: $4.00/month (92% savings)

​
What SEC API Offers That sec-api.io Doesn’t

  • Semantic search — AI-powered filing search using finance-tuned embeddings + vector search
  • Intelligence bundles — Pre-computed company, security, and earnings briefings
  • MCP integration — Native agent protocol support for Claude, Cursor, and other AI tools
  • CLI tool — npm install -g @secapi/cli for terminal workflows
  • Filing diff — Compare two filing versions to see what changed
  • Saved monitors — Webhook alerts when new filings match your criteria
  • Budget controls — Spending caps with 75/90/95% threshold alerts
  • 4 SDKs — JavaScript, Python, Go, and Rust (sec-api.io has 2)
  • 8 pre-built skills — Installable workflows for Claude Code, Codex, Cowork

​
Try It

Create an account at secapi.ai/signup to get an API key.