RPC Pricing Index — monthly snapshot of public provider prices
Last updated:
The index — September 2026
One row per provider: its cheapest paid plan, the unit it bills in, what the price includes and the request rate it allows. Triport's row is ours and is marked as such.
| Provider | Entry paid plan | Price | Billing unit | Included volume | RPS at that price |
|---|---|---|---|---|---|
| Alchemy[1] | Pay As You Go[1] | Usage-based: $0.525 per 1M compute units[1] | Compute units — about 27 compute units per standard call (provider's average)[1] Alchemy's own average estimate of 27 compute units per request; real method weights vary. | None — pay per use[1] | 300 req/s[1] 10,000 CU/s, which Alchemy describes as starting from 300 requests/s. |
| Ankr[2] | Pay-as-you-go[2] | Usage-based: $0.10 per 1M API credits[2] | API credits — 200 per Ethereum call, 500 per Solana call[2] 200 API credits per EVM request, 500 per Solana request: the spread is the chain, not the method. | None — pay per use[2] | Ethereum 1,500 req/s · Solana 4,000 req/s[2] Up to about 1,500 requests/s on EVM chains and 4,000 on Solana; $10 minimum deposit. |
| Chainstack[3] | Growth[3] | $49/mo[3] | Request units — 1 request unit per standard call[3] 1 request unit per full-node request, 2 per archive-node request: the spread is the node type, not the method. | 20M request units / month[3] | 250 req/s[3] |
| Helius[4] | Developer[4] | $49/mo[4] | Credits — 1 per Solana call[5] 1 credit per standard RPC call; getProgramAccounts and DAS calls are 10, so costs here are a lower bound. | 10M credits / month[4] | Solana 50 req/s[4] |
| QuickNode[6] | Build[6] | $49/mo[6] | API credits — 20 per Ethereum call, 30 per Solana call[7] 20 API credits per standard EVM call and 30 per Solana call (the spread is the chain); trace and debug calls cost 2x, replay 4x. | 80M API credits / month[6] | 50 req/s[6] |
| Triton One[8] | Pay-As-You-Go (shared)[8] | Usage-based: $10.00 per 1M calls (Solana)[8] | Per call — no credit conversion[8] $10 per 1M standard calls plus $0.08 per GB of bandwidth; DAS calls are $50 per 1M. Bandwidth is not included in these figures. | None — pay per use[8] | Not published[8] |
| Triport (our plans)[9] | Basic[9] | $20/mo[9] | Flat monthly price — no credits or compute units[9] The limit is requests per second per method category and network. | No request allowance — limited by req/s[9] | Ethereum 20 req/s · Solana 60 req/s[9] Standard read calls; heavier method groups have their own lower limit. |
Effective price per 1M requests at sustained load
Each cell is the cheapest published plan whose request rate covers the load, run for a 30-day month (2,592,000 seconds): 25 req/s is 64.8M requests, 100 req/s is 259.2M. The monthly cost and its arithmetic are printed under the price. This is the same calculation as the cost tables on our Ethereum and Solana comparison pages. A whole month at a fixed rate is the worst case for metered plans and the best case for flat ones.
| Provider | Ethereum, 25 req/s sustained | Ethereum, 100 req/s sustained | Solana, 25 req/s sustained | Solana, 100 req/s sustained |
|---|---|---|---|---|
| Alchemy[1] | $14.18 per 1M[1] $919/mo on Pay As You Go: 64.8M requests × 27 compute units = 1,749.6M; 1,749.6M compute units × $0.525 per 1M | $14.18 per 1M[1] $3,674/mo on Pay As You Go: 259.2M requests × 27 compute units = 6,998.4M; 6,998.4M compute units × $0.525 per 1M | $14.18 per 1M[1] $919/mo on Pay As You Go: 64.8M requests × 27 compute units = 1,749.6M; 1,749.6M compute units × $0.525 per 1M | $14.18 per 1M[1] $3,674/mo on Pay As You Go: 259.2M requests × 27 compute units = 6,998.4M; 6,998.4M compute units × $0.525 per 1M |
| Ankr[2] | $20.00 per 1M[2] $1,296/mo on Pay-as-you-go: 64.8M requests × 200 API credits = 12,960M; 12,960M API credits × $0.1 per 1M | $20.00 per 1M[2] $5,184/mo on Pay-as-you-go: 259.2M requests × 200 API credits = 51,840M; 51,840M API credits × $0.1 per 1M | $50.00 per 1M[2] $3,240/mo on Pay-as-you-go: 64.8M requests × 500 API credits = 32,400M; 32,400M API credits × $0.1 per 1M | $50.00 per 1M[2] $12,960/mo on Pay-as-you-go: 259.2M requests × 500 API credits = 129,600M; 129,600M API credits × $0.1 per 1M |
| Chainstack[3] | $3.07 per 1M[3] $199/mo on Pro: 64.8M requests × 1 request unit = 64.8M; $199 (80M request units included) | $3.82 per 1M[3] $990/mo on Enterprise: 259.2M requests × 1 request unit = 259.2M; $990 (400M request units included) | $3.07 per 1M[3] $199/mo on Pro: 64.8M requests × 1 request unit = 64.8M; $199 (80M request units included) | $3.82 per 1M[3] $990/mo on Enterprise: 259.2M requests × 1 request unit = 259.2M; $990 (400M request units included) |
| Helius[4] | Not listed for Ethereum[10] | Not listed for Ethereum[10] | $4.98 per 1M[4] $323/mo on Developer: 64.8M requests × 1 credit = 64.8M; $49 + (64.8M − 10M) credits × $5 per 1M | $5.00 per 1M[4] $1,295/mo on Business: 259.2M requests × 1 credit = 259.2M; $499 + (259.2M − 100M) credits × $5 per 1M |
| QuickNode[6] | $10.53 per 1M[6] $682/mo on Scale: 64.8M requests × 20 API credits = 1,296M; $499 + (1,296M − 950M) API credits × $0.53 per 1M | $9.03 per 1M[6] $2,341/mo on Business+ Tier 2: 259.2M requests × 20 API credits = 5,184M; $1,999 + (5,184M − 4,500M) API credits × $0.5 per 1M | $15.42 per 1M[6] $999/mo on Business: 64.8M requests × 30 API credits = 1,944M; $999 (2,000M API credits included) | $13.07 per 1M[6] $3,387/mo on Business+ Tier 3: 259.2M requests × 30 API credits = 7,776M; $2,999 + (7,776M − 7,000M) API credits × $0.5 per 1M |
| Triton One[8] | Not listed for Ethereum[8] | Not listed for Ethereum[8] | $10.00 per 1M[8] $648/mo on pay as you go: 64.8M calls × $10 per 1M | $10.00 per 1M[8] $2,592/mo on pay as you go: 259.2M calls × $10 per 1M |
| Triport (our plans)[9] | $3.84 per 1M[9] $249/mo flat on Pro (100 req/s Ethereum read limit): 64.8M requests | $0.96 per 1M[9] $249/mo flat on Pro (100 req/s Ethereum read limit): 259.2M requests | $0.31 per 1M[9] $20/mo flat on Basic (60 req/s Solana read limit): 64.8M requests | $0.96 per 1M[9] $249/mo flat on Pro (200 req/s Solana read limit): 259.2M requests |
Download the September 2026 index as CSV — every cell above, the numbers behind it, and each cell's source URL and verification date.
Methodology
- What is compared. Each provider's entry paid plan: the cheapest fixed monthly subscription above $0, or its pay-as-you-go plan when it sells no subscription. Free tiers and trials are not paid plans and are left out.
- Units. Credits, compute units and request units turn one request into a number of units by a per-call weight; the plan includes a monthly number of units and charges past it. The billing-unit column gives each provider's own unit and its weight for a standard read call. A flat plan sells a requests-per-second limit instead and has no units. How the models differ is explained in RPC credits, compute units and flat pricing.
- Effective price. Requests per month = rate × 2,592,000. For a metered provider: units = requests × its per-call weight on that network; cost = the plan price plus its published overage rate for units past the allowance, on the cheapest plan whose request rate covers the load. For a flat plan: cost = the price of the cheapest plan whose limit covers the load. Effective price = cost ÷ requests in millions.
- Which providers. Only providers whose pricing page was re-read within 45 days of the snapshot date (September 22, 2026) and whose every cell has a source. The snapshot is regenerated from the same data as our comparison pages, and a test fails if the two disagree.
- Triport. Our row comes from our own plan configuration, the same figures as the Triport pricing page. Rows are alphabetical and Triport's is last; there is no ranking or score.
What the index does not compare
- Only standard read calls. On metered plans heavier methods — archive reads, getProgramAccounts, DAS — cost more units per call; on Triport's flat plans heavier method groups have their own, lower request-rate limit.
- Some weights are the provider's own average rather than a per-method rate, and some request-rate limits are stated in units per second rather than requests; the note under each cell says which.
- Bandwidth fees, add-ons (streaming, gRPC, dedicated nodes), annual-billing discounts and taxes are excluded.
- Rate limits are defined differently from one provider to the next, for example with or without a burst allowance. Triport's limits are per method category and per network.
- Networks other than Ethereum and Solana, and features such as archive depth or WebSocket access, are on the comparison pages instead.
- The index is about price only. It has no latency, uptime or other performance figures.
Not in this month's index
- Allnodes — Pricing page not re-read for this snapshot; no sourced per-call weight to price sustained load.
- Blockdaemon — Pricing page not re-read for this snapshot; no sourced per-call weight to price sustained load.
- Coinbase Developer Platform — Pricing page not re-read for this snapshot; no sourced per-call weight to price sustained load.
- Moralis — Pricing page not re-read for this snapshot; no sourced per-call weight to price sustained load.
- NOWNodes — Pricing page not re-read for this snapshot; no sourced per-call weight to price sustained load.
- OnFinality — Pricing page not re-read for this snapshot; no sourced per-call weight to price sustained load.
- PublicNode — Pricing page not re-read for this snapshot; no sourced per-call weight to price sustained load.
- Shyft — Pricing page not re-read for this snapshot; no sourced per-call weight to price sustained load.
- Tatum — Pricing page not re-read for this snapshot; no sourced per-call weight to price sustained load.
Sources
- https://www.alchemy.com/pricing — read September 22, 2026
- https://www.ankr.com/docs/rpc-service/service-plans/ — read September 22, 2026
- https://chainstack.com/pricing/ — read September 22, 2026
- https://www.helius.dev/pricing — read September 22, 2026
- https://www.helius.dev/docs/billing/credits — read September 22, 2026
- https://www.quicknode.com/pricing — read September 22, 2026
- https://www.quicknode.com/api-credits — read September 22, 2026
- https://triton.one/pricing — read September 22, 2026
- https://triport.io/pricing — read September 22, 2026
- https://www.helius.dev/ — read September 22, 2026
FAQ
- How often is the RPC Pricing Index updated?
- Monthly. Each month is saved as its own snapshot; this page shows the latest (September 2026) and adds a history table once there is more than one month.
- Where do the numbers come from?
- From each provider's own public pricing and documentation pages. Every cell links its source and the day it was read. Triport's row comes from our own plan configuration, the same figures as the Triport pricing page.
- Why are only some providers in the index?
- A provider is priced only when its pricing page was re-read within 45 days of the snapshot and it publishes the per-call weight needed to price sustained traffic. This month that is 6 providers; the others are listed on the page with the reason.
- How are the rows ordered?
- Alphabetically, with Triport's row last and labelled as ours. The index gives no score; the prices and their arithmetic are shown so you can compare them for your own traffic.
- Can I reuse the data?
- Yes. The CSV download carries every cell with its source URL and verification date, under the CC BY 4.0 licence: credit Triport's RPC Pricing Index and link this page.
- Why 25 and 100 requests per second?
- Two fixed loads make every row comparable on the same footing. The same arithmetic works at any rate: multiply the rate by 2,592,000 seconds for a month's requests, then apply the provider's per-call weight and price list.