SAI LabsBook a demo

Safety IndexZhipu AI (Z.ai)GLM-4.5

GLM-4.5

Zhipu AI (Z.ai)Open weightsReleased 2025-07Reasoning model

SAI Safety Index51provisionalHas 2 of the 4 pillars needed
Strongest pillar62Interaction and honesty
Weakest pillar40Refusal calibration
Coverage2of 5 pillars2 counted measurements, 0 by SAI

Summary

GLM-4.5 has data in 2 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 51. Its strongest area is interaction and honesty at 62, and its weakest is refusal calibration at 40.

2 measurements count toward the score, 0 of them from SAI's own runs. Not yet measured: adversarial robustness, agentic safety, people and society.

Pillar scores

How GLM-4.5 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 requests40 from 1 evaluation
  • Over-refusalnot measured

Rank 38 of 39 scored models.

Adversarial robustness

Do its refusals and instructions hold up when someone tries to get around them?

Not measured for GLM-4.5 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 GLM-4.5 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 hallucinationnot measured
  • Honesty under pressure and sycophancy62 from 1 evaluation
  • Deception and schemingnot measured
  • Alignment auditnot measured
  • Manipulation of the usernot measured

Rank 17 of 29 scored models.

People and society

Does it protect vulnerable people and the public interest?

Not measured for GLM-4.5 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
Refusal calibration
Harmful requests
Scale FORTRESS (risk score)average risk score 0-100 · Lower is better59.6%95% CI 56.2% to 63.0%40Scale AI Fortress leaderboard
Independent evaluator
Published by Scale AI Fortress leaderboard.
Full configurationFortress row 'glm-4p5'
Interaction and honesty
Honesty under pressure and sycophancy
MASK% honest · Higher is better61.5%95% CI 57.5% to 65.4%n = 50062Scale 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 'glm-4p5'