gpt-5.2

Score

65

Task done

75.61%

59–91% 95% CI bootstrap over items · n=95 · measured 2026-10-04

Refusal quality

91.1%

names the gap 90% · offers a route 87% · 10 probes x 3 iterations

Fabrication

10%

3 of 30 probe runs

Dead-air

71.0%

245 of 345 turns

Tool silence

100.0%

all of 19 items (245 tool-call turns) · ≥82%

Stalled

28.9%

26 of 90 runs

TTFT p50

664ms

664–825 p50–p90 · n=38

Cost / 1M tok

$14.00