solana-rpc-bench
A Solana RPC benchmark you run from your own server
Compare Helius, QuickNode, Alchemy, Triton or the public endpoint from where your app runs: latency per method, errors and 429s, slot lag against the other endpoints, which heavier methods each one serves, and a WebSocket check. API keys are hidden in every output. Sequential and gentle. Zero dependencies.
Try it
npx github:blixvip/solana-rpc-bench https://api.mainnet-beta.solana.com --rounds 1 --delay 500 --ws
# several endpoints, keys included (they are redacted in the output):
SOLANA_RPC_URLS="https://a...,https://b..." npx github:blixvip/solana-rpc-bench
Example output
A real run against the public endpoint on 2026-11-17 (--quiet hides the per-request progress):
solana-rpc-bench 2026-11-17T21:23:38.431Z 1 round, 500 ms between requests, commitment confirmed
#1 https://api.mainnet-beta.solana.com
requests 9, errors 0, 429s 1, timeouts 0; HTTP 200 x8, 429 x1
slot lag vs highest seen: median 0, max 0 (1 samples, last slot 470682967)
method ok p50 ms p90 ms max ms last error
getSlot 1/1 290 290 290
getLatestBlockhash 1/1 89 89 89
getBalance 1/1 118 118 118
getAccountInfo 1/1 116 116 116
getSignaturesForAddress 1/1 95 95 95
getTransaction 1/1 101 101 101
supports:
getTokenLargestAccounts no: HTTP 429 (code 429): Too many requests for a specific RPC call
getProgramAccounts (filters) yes (297 ms)
getRecentPrioritizationFees yes (96 ms)
websocket wss://api.mainnet-beta.solana.com: open 288 ms, subscribed 379 ms, first slot 595 ms
One location, one time window. Latency includes your own network path to each provider. Not a ranking of providers.
What it measures
- p50, p90 and max for
getSlot,getLatestBlockhash,getBalance,getAccountInfo,getSignaturesForAddressandgetTransaction - Errors, HTTP status counts, 429s and timeouts
- Slot lag: every endpoint's slot read back to back, each round
- Whether it serves
getTokenLargestAccounts, filteredgetProgramAccountsandgetRecentPrioritizationFees(or your own list) - WebSocket open,
slotSubscribeand first notification (Node 22+) - Text,
--jsonor--markdownoutput
Use it in your code
npm install github:blixvip/solana-rpc-bench
import { bench, formatTable } from "solana-rpc-bench";
const report = await bench(
["https://api.mainnet-beta.solana.com", process.env.MY_RPC_URL].filter(Boolean),
{ rounds: 3, delayMs: 300, ws: false },
);
console.log(formatTable(report));
report.endpoints[0].methods.getAccountInfo.p50; // ms
report.endpoints[0].slotLag.max; // slots behind the highest seen
Run it on a schedule or share a report
- examples/github-action/bench.yml: copy into your repository's workflows to bench your endpoints daily from GitHub's runners, with the Markdown report in the job summary.
- examples/html-report.js: turn
--jsonoutput into one self-contained HTML page.
Full documentation, the API and the caveats are in the README. MIT licensed, no dependencies.
Guides
From the team behind Blix
solana-rpc-bench is a small, open piece of the tooling behind Blix, the AI-powered Solana memecoin terminal with 0% fees, which runs on several RPC keys at once: every pump.fun launch live, AI coin ratings, narratives, wallet and X tracking, and one-click trading. There is also a public API and an MCP server for AI agents. This tool never calls a Blix service.
Contribute
Pull requests are welcome and there are good first issues with step-by-step instructions. Setup is git clone and npm test, fully offline. Ask anything in Discussions or the Discord.