Tier 1 reference benchmarks
Reproducible
rankings by
scenario.
Side-by-side metrics, no composite weighting, no hidden tie-breakers. Each entry can be re-run with the same seed and will produce the same result under Tier 1 conditions.
Scenario
Sort by
| Rank | Name | Scenario | Agents | Delivery | Deal | Latency | Throughput | Date |
|---|---|---|---|---|---|---|---|---|
| 01 | Marketplace v3 (optimized) | Marketplace | 100 | 100% | 68.4% | 2.1t | 42.3 m/t | 2026-05-14 |
| 02 | Auction with dynamic pricing | Auction | 50 | 100% | 56% | 1.8t | 31.2 m/t | 2026-05-13 |
| 03 | Quorum Consensus (7 nodes) | Consensus | 7 | 100% | — | 1t | 15.6 m/t | 2026-05-12 |
| 04 | Supply chain (4-hop) | Supply chain | 4 | 100% | — | 4t | 7.8 m/t | 2026-05-11 |
| 05 | Voting with 20 voters | Voting | 22 | 100% | — | 1.2t | 22.3 m/t | 2026-05-10 |
| 06 | Reputation (20% malicious) | Reputation | 10 | 100% | — | 2.4t | 11.7 m/t | 2026-05-09 |
| 07 | Marketplace baseline | Marketplace | 20 | 100% | 52.1% | 2.8t | 9.9 m/t | 2026-05-08 |
| 08 | Consensus under partition | Consensus | 5 | 95.2% | — | 6.3t | 4.5 m/t | 2026-05-07 |
Methodology
About these
rankings.
Rankings come from Tier 1 reference runs. Re-run with the same seed and you get the same numbers. Tier 1 uses virtual tick-based time with no transport failures, so delivery rates are expected to be near 100%.
Delivery rate
Fraction of sent messages that were received. Not a measure of protocol success.
Deal rate
Percentage of buy requests that resulted in a successful trade. Marketplace and auction only.
Latency
Mean ticks between send and receive for correlated message pairs.
Throughput
Messages processed per tick across all agents.