Weaviate scales test-time compute in search, boosting nDCG to 57.5
lateinteraction · x · 2026-08-19
Weaviate introduces test-time compute scaling in Search Mode via a new effort parameter, balancing accuracy and latency. On the BRIGHT Biology benchmark, ultrahigh effort lifts nDCG@10 to 57.5 compared to 13.0 with Hybrid Search. This feature, available in weaviate-agents 1.8.0, scales computation during query decomposition and reranking.
More from Infra
- Developer Exhausts 20x Usage Limits on Both OpenAI and Anthropic — antirez · 2026-08-19
- Robotic Simulation Highlights as Clean Innovation Incentive System — const_reborn · 2026-08-19
- Users seek benchmarks on token subsidies for subscriptions like Claude and Codex Pro — mitsuhiko · 2026-08-19
- Meta Releases Muse Glimmer: 30B Model Optimized for Always-On Local Voice Agents — Once_ina_Lifetime · 2026-08-19
- Gortex Engine Indexes Code into Graphs to Cut Token Usage by 50x — tom_doerr · 2026-08-19
- A100 study shows high GPU utilization masks low tensor core efficiency — tokenbender · 2026-08-19