PRISM Leaderboard

← 2026-Q4 board

Playwright 1.60.0

PRISM Benchmark · Operator baseline

0.4599

R̄ 0.6409 · S̄ 0.6900 · Ī 0.5200 · worst case 0.4460

False Heals 8 of 50 · intent violations 2 · integrity violations 0

By category

CategoryScenariosIndexR̄S̄ĪFalse HealsIntentIntegrity
Accessibility & Shadow DOM Obstructions50.20000.40000.20000.20001 of 500
Affective & Cognitive Load50.60000.90581.00000.60000 of 500
Canvas & WebGL Blind Spot50.20000.20000.20000.20000 of 500
Dynamic Data & Network Artifacts50.70000.70000.80000.80001 of 500
Hydration & Timing Gap50.43950.63951.00000.80001 of 500
Lying DOM50.80001.00000.80000.80001 of 500
Non-Deterministic & Generative UI50.20000.30000.30000.20001 of 500
Peak Complexity & Multi-Step Logic50.00000.50000.60000.00002 of 520
React-Specific State & Race Conditions50.66000.86001.00000.80001 of 500
Visual Occlusion & Layout Shifts50.80000.90351.00000.80000 of 500

Provenance

Archive SHA-256
945c296036d2c135820daa44baa0bf5a8b0c83e80152ba65b4d4a41bd02dab26
Seed set
2026-Q4
Sandbox version
1.0.0
Generated
2026-09-21T06:24:29.000Z
Scope
full

Every quarter Playwright has entered →