One JSON-RPC Call, Nine Endpoints, Four Chains: What Public Gas Prices Actually Say Right Now
One JSON-RPC Call, Nine Endpoints, Four Chains: What Public Gas Prices Actually Say Right Now
Disclosure: this article was written with AI assistance from measurements taken by our own tooling, then reviewed for accuracy. Nothing here is financial advice — this is a measurement log, not a recommendation to use, avoid, or move funds to any chain.
#ad — affiliate note: this article currently contains no affiliate or sponsored links. If that ever changes, affiliate links will be marked here and labeled clearly.
Every comparison of "which chain is cheaper" eventually rests on one number: the gas price. And every one of those numbers is quoted from someone else's readout. We wanted to know what the raw, first-hand data looks like when you skip the dashboards and ask the chains directly — so we made one standard JSON-RPC call, eth_gasPrice, to four EVM chains through nine public endpoints run by independent operators, twice, ninety seconds apart. What we found is that the prices agree remarkably well inside each chain, several sit on hard floors, and one common assumption did not survive contact with the data.
The method (repeatable in five minutes)
eth_gasPrice is the standard JSON-RPC method every EVM node exposes for fee suggestions. It needs no account and no API key. The exact request we sent to each endpoint, four chains, two rounds:
POST <rpc-url> Content-Type: application/json {"jsonrpc":"2.0","id":1,"method":"eth_gasPrice","params":[]} The response is a hex quantity in wei. The endpoints, chosen so each chain has at least two operators with no shared infrastructure:
- Ethereum mainnet:
ethereum-rpc.publicnode.com,eth.drpc.org,1rpc.io/eth - BNB Smart Chain:
bsc-dataseed1.binance.org,bsc-mainnet.public.blastapi.io - Arbitrum One:
arb1.arbitrum.io/rpc(the chain's own official endpoint) andarbitrum-one-rpc.publicnode.com - Optimism (OP Mainnet):
mainnet.optimism.io(official) andoptimism-rpc.publicnode.com
Round 1 ran at 02:03:42 UTC and round 2 at 02:05:24 UTC on 2026-11-03. Ten requests total; one (Arbitrum's publicnode endpoint, round 2) failed with a connection reset, so we report nine completed readings. All values below are snapshots that will be wrong by the time you read this — the structure is the point, and you can re-run the probes yourself.
The readings
Values converted from wei; gwei = one-billionth of an ETH/BNB/ARB token unit.
| Chain | Round 1 (gwei) | Round 2 (gwei) | Agreement across providers |
|---|---|---|---|
| Ethereum | 0.051654 / 0.051654 / 0.051651 | 0.040292 / 0.040292 / 0.040292 | Byte-identical hex, all three, both rounds |
| BSC | 0.050000 / 0.050000 | 0.050000 / 0.050000 | Identical all four calls |
| Arbitrum One | 0.020052 / 0.020048 | 0.020000 | Within 0.3% |
| Optimism | 0.001000 / 0.001000 | 0.001000 / 0.001000 | Byte-identical hex, all four calls |
The Ethereum agreement is the striking one. On round 1, publicnode and drpc returned exactly 0x3142e03 — not "roughly the same number," the same 54,048,771 wei — while 1rpc was within 0.006%. On round 2, all three independent operators returned the identical hex 0x266ccbe, and mainnet had meanwhile fallen 22% (from 0.0517 to 0.0403 gwei) inside 102 seconds. Independent providers converge because they read the same pending-transaction landscape and, at today's fee levels, everyone's suggestion logic lands on the same floor.
Three floors, and one assumption that broke
Look at BSC: four calls to two independent providers, and every single one returned exactly 0x2faf080 = 0.05 gwei. A number that pinned cannot be a market price — it is a minimum accepted gas price, a floor below which the chain will not accept transactions. Optimism shows the same signature at exactly 0.001 gwei (hex 0xf4364, one-tick variation between rounds). Arbitrum hovers at 0.02 gwei. These endpoints are telling you the chains' rate limits, not their congestion.
Which brings us to the assumption that broke. If you asked most people to rank "mainnet vs the BNB chain" on gas price, they would not put Ethereum first. Yet in this window, Ethereum mainnet's suggested gas price (0.0403 gwei on round 2) came in below BSC's floor (0.05). Read that sentence precisely, because the temptation to over-interpret it is exactly the trap this article exists to describe.
What gas price is not
A gas price is one factor of a three-factor invoice, and none of the readings above tells you what an operation costs:
- Units differ per operation and per chain. A plain ETH transfer burns a fixed 21,000 gas, so mainnet cost ≈ gas price × 21,000. Other operations — swaps, mints, deposits — multiply by tens or hundreds of times more gas, and the exact figure depends on the contract and state access, not the chain's headline rate.
- On the two Layer 2s above,
eth_gasPricecaptures only the execution leg. Arbitrum and Optimism additionally charge for posting your transaction data to Ethereum (Arbitrum's two-dimensional fee model; Optimism's L1 cost component). A low L2 gas price multiplied by a small execution unit still has to be added to an L1 data fee that this method does not report at all. - The token price is the third multiplier. 0.04 gwei priced in ETH and 0.001 gwei priced in OP are only comparable after you multiply by two different market prices.
So the honest ratios this window supports are gas-price-term ratios only: Ethereum's suggested gas price sat about 40x above Optimism's, roughly 2x above Arbitrum's, and just under 1x of BSC's floor. Whether any given transfer actually costs less on Optimism than on BSC depends on units, the L1 fee, and token prices — numbers we deliberately did not convert here, because we measured the inputs, not your invoice.
Why asking the chains directly is worth the five minutes
Every number in this article came from the chain operators themselves — two of the nine endpoints are the chains' own official public RPCs — with no aggregator, no dashboard, and no account in between. When three operators that share nothing return the identical hex quantity, you have a fact about chain state, not a product's opinion. When a value freezes at the same hex across providers and minutes, you have found a floor. Both are things you can only see by asking directly.
Re-run the probe any day you like: the request body is above, all endpoints are public, and no key is required. Note the UTC timestamp on your readings, because on Ethereum, gas prices we logged moved 22% inside under two minutes.
Not financial advice. Fee data changes block by block; nothing here predicts future costs or recommends any chain, token, RPC provider, or exchange.
Verify official sources. Read the fee mechanics yourself at the primary sources: the official Arbitrum and Optimism documentation on how L2 fees are composed (arbitrum.io and docs.optimism.io), each RPC operator's own site for its endpoint terms, and your wallet or node for your actual transaction's gas usage. The readings in this article were captured live on 2026-11-03 at 02:03–02:05 UTC directly from the listed endpoints and can be re-measured by anyone, without an account.
FeeAtlas publishes measurement logs of exchange and network fees, written with AI assistance, from first-hand data only.