HyperIndex Performance Benchmarks
HyperIndex is the fastest blockchain indexer in four of the five scenarios in our open benchmark, jointly in one of them, and second in the fifth. On state aggregation it processed 261x the events per second of a subgraph on Graph Node.
Figures are events per second from the run of 16 September 2026. The benchmark runs in GitHub CI and checks each tool's output against ground truth. The repository README always has the latest run, and the methodology explains how it works.
State aggregation
Every rETH transfer changes a balance, so each event means finding a row, updating it and saving it again.
| Tool | Data source | Events/s | vs fastest |
|---|---|---|---|
| Envio Indexer | HyperSync | 8,182.4 | Fastest |
| Envio Subgraph | HyperSync | 7,649.6 | 1.1x slower |
| Rindexer | HyperSync | 947.2 | 8.6x slower |
| Squid SDK | SQD Network | 556.1 | 14.7x slower |
| Envio Indexer | RPC | 341.9 | 23.9x slower |
| Rindexer | RPC | 322.8 | 25.3x slower |
| Envio Subgraph | RPC | 241.1 | 33.9x slower |
| Ponder | RPC | 62.0 | 132x slower |
| Subgraph (Graph Node) | RPC | 31.4 | 261x slower |
| Substreams | StreamingFast | 28.7 | 285.1x slower |
| SubQuery | RPC | 25.3 | 323.9x slower |
| Squid SDK | RPC | 17.3 | 474.1x slower |
SubQuery and Squid SDK on RPC did not finish verifying the full range inside the 300 second cap.
Solana USDC transfers
Every USDC transfer on Solana, including the ones inside swaps and routers.
| Tool | Data source | Events/s | vs fastest |
|---|---|---|---|
| Envio Indexer | HyperSync | 16,204.7 | Fastest |
| Substreams | StreamingFast | 3,856.5 | 4.2x slower |
| Squid SDK | SQD Network | 996.8 | 16.3x slower |
| Carbon | RPC | 563.2 | 28.8x slower |
Other scenarios
| Scenario | Fastest | HyperIndex | Next tool |
|---|---|---|---|
| Decoded event stream | Rindexer on HyperSync, 78,416 | 1.2x slower | Squid SDK on SQD Network, 6.7x slower |
| External contract calls | HyperIndex and Squid SDK, jointly | Joint fastest | Rindexer on HyperSync, 1.7x slower |
| Factory contract registration | HyperIndex, 10,782 | Fastest | Rindexer on HyperSync, 1.5x slower |
Rindexer's lead on the decoded event stream comes from reading HyperSync, Envio's own data layer. With RPC as its source, Rindexer is 7.2x slower, at 10,950 events/s. Read more about HyperSync.
For a wider comparison, see Best Blockchain Indexers in 2026.
Sentio benchmark cases, April 2025
These cases come from Sentio's April 2025 research and are kept for reference. They measure total sync time, predate the current method and are not comparable with the tables above. The Envio column names the product behind each figure, because HyperSync is a raw data engine rather than a full indexer. Sentio has since published its own benchmark with different results.
| Case | Description | Envio | Nearest competitor | The Graph | Ponder |
|---|---|---|---|---|---|
| LBTC Token Transfer Events | Event handling, no RPC calls, write-only | 3m (HyperIndex) | 8m - 2.6x slower (Sentio) | 3h9m - 63x slower | 1h40m - 33x slower |
| LBTC Token with RPC calls | Event handling, RPC calls, read-after-write, point calculation | 1m (HyperIndex) | 6m - 6x slower (Sentio) | 1h3m - 63x slower | 45m - 45x slower |
| Ethereum Block Processing | 100K blocks with metadata extraction | 7.9s (HyperSync) | 1m - 7.5x slower (Subsquid) | 10m - 75x slower | 33m - 250x slower |
| Ethereum Transaction Gas Usage | Transaction handling, gas calculations | 1m26s (HyperSync) | 7m - 4.8x slower (Subsquid) | N/A | 33m - 23x slower |
| Uniswap V2 Swap Trace Analysis | Transaction trace handling, swap decoding | 41s (HyperSync) | 2m - 2.9x slower (Subsquid) | 8m - 11x slower | N/A |
| Uniswap V2 Template | Event handling, pair and swap analysis | 8s (HyperIndex) | 2m - 15x slower (Subsquid) | 19m - 142x slower | 21m - 157x slower |
Some runs in these cases did not return complete data. The reference README has the counts.
Historical benchmarking results
In October 2023 our internal benchmark indexed the Uniswap V3 ETH-USDC pool 2.1x faster than the nearest competitor. See the Indexer Benchmarking Results blog post.
Verify for yourself
Every Open Indexer Benchmark figure comes from code you can run. Each scenario in the repository has its own setup instructions, and contributions are welcome, including new indexers, new scenarios and reports of results that look wrong.
