Last updated: October 1, 20267 min read

Stale prices on a swap, wallet balances that won't update, historical scans that time out right when a customer needs them are not necessarily app-level issues. Often, it's the RPC connection underneath it that is the bottleneck.

Solana RPC is the connection between a Solana app and the chain itself: the interface a backend or client calls to read account balances, program state, and transaction history, and to submit transactions. Every piece of data an app displays relies on this connection, so when the connection is slow, your app is slower - causing traders to lose money and frustrating users.

Why Solana RPC read performance matters

Solana was built for speed: blocks about every 270 milliseconds as of this writing, still coming down toward 200, plus thousands of transactions per second and sub-cent fees. That speed sets the bar for everything built on top of it. If an RPC provider adds fifty, a hundred, or three hundred milliseconds to every read, the app is only as fast as that provider's queue.

From speaking with our customers, here are some examples of common pain points:

A DEX or aggregator quotes a stale price, and a swap executes at a worse price and loses money.
A wallet's balance view lags behind the chain, so a user can't tell if their deposit landed.
An indexer's backfill takes weeks instead of days, delaying every feature built on top of it.
A compliance or tax tool times out scanning a wallet's full history at the exact moment a customer needs the answer.

Reads, not writes, are where most Solana apps spend their infrastructure budget. Transaction landing (submitting and confirming a transaction) is an estimated 5% or less of total Solana RPC traffic across the ecosystem. The rest is reads: balances, account state, program data, and historical lookups. That's why read latency deserves just as much attention as landing speed.

What the benchmark data shows

We publicly publish our Solana RPC benchmark, testing every provider from the same AWS regions (US East, US West, EU Central, AP Southeast) with the same methods, on a rolling 24-hour window that refreshes every five minutes, against four other providers: QuickNode, Chainstack, Helius, and Triton. Here's what the current window shows:

Alchemy has the lowest average Solana read latency globally: 9.21 ms, about 26% faster than the next-closest provider (QuickNode) and roughly 11x faster than the slowest provider we track (Triton).
On P95 tail latency, Alchemy is ~45% faster than the next-closest provider and roughly 20x faster than the slowest.
Success rates are close to parity across all five providers (99.99–100% in this window), but Alchemy has proven historical reliability under load, providing 99.99% uptime even during the largest crypto liquidation event in history.

Full methodology, every region, and live numbers refreshing every five minutes are public on the Solana RPC benchmark.

Architecting Solana RPC for industry-leading speed and reliability

Alchemy's Solana infrastructure is purpose-built by dedicated Solana engineers, layered on top of the routing, proxying, and node management systems Alchemy has run across 100+ chains for 9+ years.

A few concrete pieces of that architecture, as detailed on Alchemy's Solana page:

An edge proxy sits in front of every request and handles connection reuse, auth, and routing before a call reaches a node, cutting tail latency by up to 7.5x in Alchemy's internal testing.
Contextual smart routing sends each request to the node best positioned to answer it fastest, instead of round-robining across a fixed pool, for up to 2.5x faster queries in Alchemy's internal testing.
On Solana, internal testing shows that architecture delivers 10x faster performance on heavy methods like getProgramAccounts and 20x faster archival data access. It's backed by 99.99% uptime, multi-region infrastructure, 48-hour block replay, and three to five layers of autonomous failover.
Staked connections for Solana specifically improve transaction landing rate on top of the read-side gains.

It's edge routing, node operations, regional placement, and archival architecture working together, which is why the Solana benchmark tests the full request path end to end instead of one component in isolation.

Exclusive Solana methods Alchemy offers

Beyond making the standard Solana RPC methods faster, Alchemy has shipped methods that don't exist anywhere else, solving problems that used to require builders to run their own indexing or account archival infrastructure.

Method

What it unlocks

getAccountInfo with slot, lastUpdateBeforeSlot, or firstUpdateAfterSlot (Solana Account Archive)

Query any account's exact state at any historical slot. A Solana validator only ever stores an account's latest state; this rebuilds the history that's otherwise gone the moment an account is touched.

getTokenAccountsByOwnerAtSlot

Return a wallet's full token holdings, with exact balances, as of any past slot in one call, instead of replaying thousands of transactions to reconstruct it.

getTokenLargestAccounts, enhanced

Return 1,000+ token accounts per call, 50x more than the standard limit.

gRPC replay slots

Recover up to 48 hours of historical slot data on demand, so a stream interruption doesn't mean lost data.

In Alchemy's internal testing, getTokenAccountsByOwnerAtSlot results matched the real balances recorded on Solana exactly, including deep history like a token balance at slot 80,000,000. The Solana Account Archive serves historical reads with median latency in the microseconds and never deletes old data, so it can answer queries reaching further back in history over time.

For teams doing backtesting, audits, incident recovery, tax and cost-basis reporting, or portfolio history, these methods replace what used to be a custom indexing project with one RPC call.

What are the key use cases for fast Solana reads?

Wallets and portfolio apps. Balance checks, token account lookups, and swap or transfer history all run through reads, and users notice lag immediately. This is the exact workload Solflare measured when evaluating archival infrastructure for its analytics features (more below).

DEXs, aggregators, and trading tools. Price and pool state reads feed directly into what a user sees before they trade. A stale getAccountInfo or getProgramAccounts call means a stale quote, and on a chain producing a block about every 270 milliseconds, stale can mean wrong within a single block.

Indexers and analytics platforms. Backfilling months or years of Solana history means running heavy, filter-heavy calls like getProgramAccounts and getTokenAccountsByOwner at scale, plus archival getTransaction and getBlock calls that stay fast without rebuilding from genesis every time.

Payments, fintech, and remittance apps. Deposit and balance confirmation needs to be fast and consistent, especially when a user is staring at a screen waiting to see funds land.

Compliance, tax, and reporting tools. Full-history scans across large numbers of accounts depend on archival reads that stay fast at the tail end of a wallet's history, not just on the most recent transactions.

Real-time trading and market data. For workloads where every millisecond of a slot update matters, gRPC streaming (Yellowstone-compatible on Alchemy) pushes account and transaction updates as they happen instead of polling for them. Alchemy connects to multiple upstream nodes, so one node's outage or restart doesn't interrupt the stream.

What proof is there that this holds up in production?

Many of the leading teams building on Solana — Phantom, Solflare, Collector Crypt, Slingshot, and others — run on Alchemy specifically for this kind of read performance and reliability. Here are a few of their stories.

, the leading multichain, self-custodial wallet trusted by 20M+ users, has been an Alchemy customer since 2022.
Solflare, one of the largest and most trusted wallets in the Solana ecosystem with 4M+ active users, uses our Solana Archive Method for up to 20x faster archive queries and full wallet history backfills in under a minute.
Collector Crypt, the leading platform for tokenized collectibles with nearly $40M in trading card inventory onchain, routes 29M+ requests through Alchemy every 30 days on its settlement and confirmation path, averaging under 7ms on common Solana read methods.
Cielo, an onchain activity tracker that lets users follow wallets and tokens in real time, runs on Alchemy across Bitcoin, Solana, Sui, Tron, and EVM chains.

"Their infrastructure and product suite has a proven track record for performance benefits. This will be a game changer for Phantom and any other Solana developers who choose to start using Alchemy." — Francesco Agosti, CTO and Co-Founder, Phantom

"Their infrastructure and product suite has a proven track record for performance benefits. This will be a game changer for Phantom and any other Solana developers who choose to start using Alchemy."

— Francesco Agosti, CTO and Co-Founder, Phantom

How do you get started building on Solana with Alchemy?

Need real-time streaming? Point your existing Yellowstone-compatible client at Alchemy's Solana gRPC endpoint.
Want to see the numbers for yourself? Leverage the benchmarks and dive into our methodology.

We're here to help all Solana builders succeed. Apply for up to $25k in credits through the Solana Fund or reach out to chat with us.

FAQs

What is a Solana RPC provider?

A Solana RPC provider runs the node infrastructure an app calls to read blockchain state (balances, account data, transaction history) and submit transactions. Most teams don't run their own Solana nodes; they call a provider's RPC endpoint instead.

Why does RPC latency matter more on Solana than on other chains?

Solana produces a block about every 270 milliseconds as of this writing, and that interval is still coming down toward 200 milliseconds. A slow RPC read can return state that's already a block or more out of date, which matters for pricing, trading, and any feature that shows users what's happening onchain right now.

Is Alchemy's Solana RPC benchmark trustworthy?

The methodology, region list, and raw numbers are public on Alchemy's Solana benchmark page and refresh every five minutes, so anyone can check the current numbers.

What's the difference between Solana RPC and Solana gRPC?

RPC is request and response: an app asks a question and gets an answer back. gRPC is a persistent stream: a client subscribes once and receives pushed updates as they happen, which is built for real-time use cases like trading and live account monitoring.

Do I need to rewrite anything to get faster archival queries?

No. Alchemy's archival performance improvements apply to standard methods like getTransaction and getBlock with no code changes or proprietary parameters required.

What if I'm already using another Solana RPC provider?

Alchemy supports the standard JSON-RPC and Yellowstone gRPC interfaces most Solana tooling already expects, so switching providers usually means updating an endpoint URL, not changing your code.

Is there a free tier for testing?

Yes. Alchemy's free tier includes a private Solana RPC endpoint with higher limits than the public endpoint, and new Solana projects can apply for credits through the Solana Fund.