SAI LabsBook a demo

Safety IndexOpenAIGPT-6 Astra

GPT-6 Astra

OpenAIProprietaryReleased 2026-09Reasoning modelVariant scored: max

SAI Safety Index50provisionalHas 1 of the 4 pillars needed
Strongest pillar50Interaction and honesty
Weakest pillar–No data
Coverage1of 5 pillars1 counted measurements, 0 by SAI

Summary

GPT-6 Astra 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 50. Its strongest area is interaction and honesty at 50.

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-6 Astra 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?

Not measured for GPT-6 Astra yet.
  • 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?

Not measured for GPT-6 Astra yet.
  • Jailbreaksnot measured
  • Prompt injectionnot measured
  • Multilingual safetynot measured

9 models scored.

Agentic safety

Does it behave safely when it has tools and autonomy?

Not measured for GPT-6 Astra yet.
  • 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 hallucination50 from 1 evaluation
  • Honesty under pressure and sycophancynot measured
  • Deception and schemingnot measured
  • Alignment auditnot measured
  • Manipulation of the usernot measured

Rank 21 of 29 scored models.

People and society

Does it protect vulnerable people and the public interest?

Not measured for GPT-6 Astra yet.
  • 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.

EvaluationValueScoreSourceSetting
Interaction and honesty
Factuality and hallucination
AA-Omniscience (public subset)% hallucination rate · Lower is better50%n = 6,00050Artificial Analysis
Independent evaluator · 2026-10-28
Published by Artificial Analysis (2026-10-28).
Full configurationArtificial Analysis, 3,000 questions; AA-Omniscience v1.1; row 'GPT-6 Astra (max)'
Other evaluators' scores (not counted)
Vectara HHEM hallucination rate% · Lower is better8.7%context, not countedVectara 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-6 Astra'