Our January 2026 benchmark suite represents the most comprehensive performance analysis we've published. Testing was conducted on AWS c8g.metal-48xl instances with AWS Graviton4 (Neoverse V2) processors, measuring production-representative workloads across all authentication modes.
Test Infrastructure
Benchmark Environment
Instance: AWS c8g.metal-48xl
CPU: AWS Graviton4 (Neoverse V2, 96 cores)
Memory: 377 GiB DDR5
OS: Amazon Linux 2023
H33 Version: 2.4.0
All benchmarks were run with warm caches where applicable, representing typical production conditions. Cold-start measurements are noted separately.
Full Stack Authentication
Full Stack Auth combines biometric verification, FHE-encrypted matching, ZK proof generation, and optional blockchain attestation into a single call.
| Mode | Latency | Description |
|---|---|---|
| Turbo | 1.28ms | Optimized for speed, full security |
| Standard | 633µs | Balanced performance and features |
| Precision | 2.1ms | Maximum accuracy, extended checks |
Session Management
Session operations show the benefit of our context caching architecture:
| Operation | Latency | Speedup |
|---|---|---|
| Session Resume | 50µs | 4.4x vs Full Auth |
| Incremental Auth (5% delta) | <50µs | 4.4x vs Full Auth |
| Session Validation | 12µs | 18x vs Full Auth |
Proof Operations
ZK proof performance across generation, verification, and caching:
| Operation | Latency | Notes |
|---|---|---|
| Proof Generation | 1.28ms | Dilithium3 signatures |
| Proof Verification (cold) | 2.14ms | First verification |
| Proof Verification (cached) | 32µs | 67x speedup |
| Biometric ZK Proof | 260µs | FHE + ZK combined |
Batch Processing
Batch operations demonstrate sub-linear scaling for high-throughput scenarios:
| Batch Size | Total Latency | Per-User |
|---|---|---|
| 10 users | 12µs | 1.2µs |
| 100 users | 45µs | 0.45µs |
| 1,000 users | 116µs | 0.116µs |
| 10,000 users | 890µs | 0.089µs |
At 1,000 users in 116µs, that's high-throughput authentication on a single node.
Batch ZKP
| Batch Size | Total Latency | vs Sequential |
|---|---|---|
| 10 proofs | 4.2ms | 64% faster |
| 100 proofs | 35ms | 73% faster |
| 1,000 proofs | 310ms | 77% faster |
FHE Operations
Fully Homomorphic Encryption performance for biometric matching:
| Operation | Latency |
|---|---|
| Template Encryption | 85µs |
| Encrypted Matching | 260µs |
| Result Decryption | 45µs |
| End-to-End FHE Auth | 260µs |
Memory and CPU Utilization
Resource consumption under sustained load:
- Peak memory: 2.8GB for 10,000 concurrent sessions
- CPU utilization: 65% at 1M auth/sec sustained
- Cache hit rate: 94% for returning users
- GC pause: <1ms (Rust core, no GC)
Methodology
All benchmarks follow these principles:
- Warm cache: Tests run after cache warmup unless measuring cold performance
- P50 latency: Reported numbers are median (50th percentile)
- Production workloads: Test data represents real authentication patterns
- Isolated measurement: Benchmarks run in isolation to avoid interference
- Repeated trials: Each measurement is the median of 1,000+ runs
Reproducing These Results
You can reproduce these benchmarks with your own H33 API key:
npm install @h33/benchmark-suite
h33-benchmark --api-key YOUR_KEY --suite fullThe benchmark suite includes all tests documented here and outputs comparable metrics for your infrastructure.
Run Your Own Benchmarks
Get an API key and see these performance numbers on your infrastructure.
Get Free API Key