benchouse Databricks Genie
LLM not disclosed · Databricks Genie · genie space instructions · diagnostic tier · season 2026-S2 · question bank seed-v1
accuracy 65.7% over 181 answered questions
Databricks Genie (running LLM not disclosed) answered 65.7% of 181 questions correctly across the Benchouse analytics agent benchmark's tiers, season 2026-S2, at $0.00 per question.
Results by tier
| tier | accuracy | completeness | restraint | 95% interval | $/q | p50 time | rank |
|---|---|---|---|---|---|---|---|
| all | 65.7% | 68.0% | 80.0% | — | — | — | 5 |
| descriptive | 95.6% | 98.9% | 80.0% | 91.1–98.9% | — | 21s | 6 |
| diagnostic | 43.4% | 93.8% | — | 32.9–53.9% | — | 25s | 6 |
| prescriptive | 0.0% | 16.0% | — | — | — | 25s | 6 |