Semantic Caching on SemCacheSearchQueries
6,345.61Cumulative End-to-End Inference Latency (minutes)MVR-cache
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| MVR-cache2026.05 | 6,345.61 | — | — | 111.26 | |
| vCache2026.05 | 6,361.52 | — | — | 69.77 | |
| ColBert2026.05 | 6,521.89 | — | — | 130 | |
| POQD2026.05 | 6,990.08 | — | — | 628.33 | |
| BaselinePolicy=GPTCache-style, Threshold=Pareto optimal threshold t*2026.02 | — | 2.2 | — | — | |
| KritesPolicy=Krites, Mechanism=Asynchronous VerifyAndPromote loop, Threshold=Pareto optimal threshold t*2026.02 | — | 8.6 | 290.3 | — |