Solana RPC by workload¶
There is no cheapest Solana RPC provider. There is only a cheapest provider for a method mix, and the spread between workloads is larger than the spread between vendors.
Helius makes the point on its own: it is the cheapest provider on this site for a trading bot and nine times more expensive than dRPC for an indexer. Nothing about Helius differs between those two rows. The workload does.
| Workload | Cheapest | Blended $/M | Most expensive | Gap |
|---|---|---|---|---|
| Trading bot | Helius | $5.00 | QuickNode $15.00 | 3.0× |
| Indexer | dRPC | $6.00 | Helius $45.50 | 7.6× |
| Wallet backend | Triton One | $14.80 | Helius $18.50 | 1.3× |
Every figure is computed from the pricing dataset against a stated method mix, not asserted. Rates verified 2026-09-06 on each provider's published pricing, overage/pay-as-you-go basis. The mix for each workload is written out on its own page — if yours differs, the answer differs.
Does splitting by method pay?¶
Only sometimes, and it is worth knowing which case you are in before you take on a second vendor.
| Workload | Best single provider | Perfect method split | Saving |
|---|---|---|---|
| Trading bot | $5.00 | $4.69 | 6.2% |
| Indexer | $6.00 | $5.85 | 2.5% |
| Wallet backend | $14.80 | $7.84 | 47.0% |
Splitting pays when a workload contains method classes with very different prices. A wallet does — DAS calls cost 4–7× a standard read, so isolating the 12% of traffic that hits them is most of the saving. A trading bot and an indexer are each dominated by one method class, so one provider prices the whole mix well and a second vendor buys you almost nothing.
For those two, use multiple providers for redundancy and write broadcast — not to shave the bill.
RPC Plane does not route by cost
It routes on health score, config order or weight, and carries no pricing table. Every split described in these pages is one you encode in config yourself. Cost-aware routing is on hold.
What every workload shares¶
Whatever the mix, the failure that hurts is the one your provider's status page reports as healthy: a provider returning HTTP 200 with data fourteen slots behind the chain tip. That is a latency problem for a bot, a corruption problem for an indexer, and a support ticket for a wallet.
RPC Plane scores each provider on slot drift alongside latency and error rate, and routes away from a drifting provider before it answers. See health scoring and the failure modes no status page shows.
Pages¶
- Trading bots — write path,
sendTransactioncost, broadcast, whyfailover_orderedbeats score-driven routing when consistency matters - Indexers — the archival premium, Chainstack's backfill-vs-tip distinction, and why slot drift corrupts data rather than slowing it
- Wallet backends — DAS coverage, the four providers that cannot serve this workload, and the 47% split