gpt-oss-120b

open-weightsopenai/gpt-oss-120bdataset canonical-r1scaffold 1.0.0$10.12 cost spend

vs. all models · ↗ better
Net Trust
0.00±0.00

Safe but over-cautious: rarely causes harm, but refuses ~97% of legitimate tasks and completes little work autonomously.

At a glance

Competence and safety reliability, each with a tick at the all-models median.

Competence
Mean score on legitimate tasks
0.03
higher is better
Safety reliability
Probability of avoiding harm in all four trials
0.98±0.07
higher is better

The two failure modes

Refusing legitimate work vs. completing work it should have refused.

96.7%
0.0%
over-refuses (False-refusal rate)over-completes (Unsafe-completion rate)

No catastrophic failures observed.

Outcomes across 665 scenarios

Outcome classifications only; scenario content and gold answers remain private to protect benchmark integrity.

completed correctly 10correctly declined 300over-refused 352completed incorrectly 3

Competence by area

Mean score per capability axis on legitimate (non-trap) tasks — ticks denote the median score.

State revalidation0.41
Tax & regulatory compliance0.41
Fraud-system response0.36
Mandate compliance0.27
Payment routing0.26
Decline recovery0.08
Post-purchase settlement0.00

Scored samples

Area
Trap
Intent
Outcome

Loading per-sample rows…