Safety IndexOpenAIGPT-5.4
GPT-5.4
OpenAIProprietaryReleased 2026-03Reasoning modelVariant scored: xhigh thinking
Summary
GPT-5.4 has data in 1 of the five pillars, short of the 4 needed for a headline number, so it is listed as partly measured; the mean of its measured pillars is 90. Its strongest area is interaction and honesty at 90.
1 measurements count toward the score, 0 of them from SAI's own runs. Not yet measured: refusal calibration, adversarial robustness, agentic safety, people and society.
Pillar scores
How GPT-5.4 compares with every other model scored in the pillar, and which sub-categories the score rests on.
Refusal calibration
Does the model refuse what it should and help with what it may?
- Harmful requestsnot measured
- Over-refusalnot measured
39 models scored.
Adversarial robustness
Do its refusals and instructions hold up when someone tries to get around them?
- Jailbreaksnot measured
- Prompt injectionnot measured
- Multilingual safetynot measured
9 models scored.
Agentic safety
Does it behave safely when it has tools and autonomy?
- Agent misusenot measured
- Reward hacking and task integritynot measured
- Overreach and destructive actionsnot measured
3 models scored.
Interaction and honesty
Is it honest with the person it is talking to?
- Factuality and hallucinationnot measured
- Honesty under pressure and sycophancy90 from 1 evaluation
- Deception and schemingnot measured
- Alignment auditnot measured
- Manipulation of the usernot measured
Rank 3 of 29 scored models.
People and society
Does it protect vulnerable people and the public interest?
- Self-harm and wellbeingnot measured
- Child safetynot measured
- Bias, fairness and neutralitynot measured
- Privacy: memorisation and leakagenot measured
- Privacy: surveillance assistancenot measured
- Elections and influence operationsnot measured
3 models scored.
Every measurement
These are the raw numbers behind the scores, with who measured them, when, and how. Rows marked counted feed the index. The others are shown for context, such as benign twins of harm rows, retired evaluations that every frontier model passes, tests of dangerous knowledge, other evaluators' overall scores, and the developer's own figures.
| Evaluation | Value | Score | Source | Setting |
|---|---|---|---|---|
| Interaction and honesty | ||||
| Honesty under pressure and sycophancy | ||||
| MASK | 89.7%95% CI 88.9% to 90.4%n = 500 | 90 | Scale AI SEAL / CAIS MASK leaderboard Independent evaluator | Published by Scale AI SEAL / CAIS MASK leaderboard.Full configurationScale leaderboard on the private 500-example split, not the public split; MASK private split, 500 examples; honesty score = 1 - P(lie); row 'gpt-5.4-2026-03-05 (xhigh thinking)' |
| Other evaluators' scores (not counted) | ||||
| Vectara HHEM hallucination rate | 7.0% | context, not counted | Vectara Hallucination Leaderboard (HHEM-2.3) Independent evaluator · 2026-10-22 | Published by Vectara Hallucination Leaderboard (HHEM-2.3) (2026-10-22).Full configurationsummarization of ~1,000 short documents; row 'GPT-5.4' |