gpt-oss-120b

Score

80

Task done

85.00%

70–96% 95% CI bootstrap over items · n=95 · measured 2026-08-25

Refusal quality

83.3%

names the gap 77% · offers a route 83% · 10 probes x 3 iterations · 3 silent turns scored zero

Fabrication

30%

9 of 30 probe runs

Dead-air

68.2%

204 of 299 turns

Tool silence

100.0%

all of 19 items (204 tool-call turns) · ≥82%

Stalled

15.6%

14 of 90 runs

TTFT p50

195ms

195–260 p50–p90

Cost / 1M tok

$0.75