SAI LabsBook a demo

Safety IndexAnthropicClaude Sonnet 5

Claude Sonnet 5

AnthropicProprietaryReleased 2026-06Reasoning model

SAI Safety Index–provisionalHas 0 of the 4 pillars needed
Strongest pillar–No data
Weakest pillar–No data
Coverage0of 5 pillars0 counted measurements, 0 by SAI

Summary

Claude Sonnet 5 has no counted measurements yet. The page will fill in as runs finish.

Pillar scores

How Claude Sonnet 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?

Not measured for Claude Sonnet 5 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 Claude Sonnet 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 Claude Sonnet 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?

Not measured for Claude Sonnet 5 yet.
  • Factuality and hallucinationnot measured
  • Honesty under pressure and sycophancynot measured
  • Deception and schemingnot measured
  • Alignment auditnot measured
  • Manipulation of the usernot measured

29 models scored.

People and society

Does it protect vulnerable people and the public interest?

Not measured for Claude Sonnet 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
Other evaluators' scores (not counted)
Phare average safety% · Higher is better78.1%context, not countedGiskard Phare leaderboard
Independent evaluator · 2026-07-09
Published by Giskard Phare leaderboard (2026-07-09).
Full configurationPhare row 'Claude 5 Sonnet'
Phare bias resistance% · Higher is better39.9%context, not countedGiskard Phare leaderboard
Independent evaluator · 2026-07-09
Published by Giskard Phare leaderboard (2026-07-09).
Full configurationPhare row 'Claude 5 Sonnet'
Phare hallucination resistance% · Higher is better86.1%context, not countedGiskard Phare leaderboard
Independent evaluator · 2026-07-09
Published by Giskard Phare leaderboard (2026-07-09).
Full configurationPhare row 'Claude 5 Sonnet'
Phare harmfulness resistance% · Higher is better99.6%context, not countedGiskard Phare leaderboard
Independent evaluator · 2026-07-09
Published by Giskard Phare leaderboard (2026-07-09).
Full configurationPhare row 'Claude 5 Sonnet'
Phare jailbreak resistance% · Higher is better86.8%context, not countedGiskard Phare leaderboard
Independent evaluator · 2026-07-09
Published by Giskard Phare leaderboard (2026-07-09).
Full configurationPhare row 'Claude 5 Sonnet'
Developer-reported (not counted)
BBQ ambiguous accuracy (lab-reported)% · Higher is better98.6%developer-reported, not countedAnthropic (Claude Opus 5 system card)
Developer-reported · 2026-07-24
Reported by the model's developer in its own system card or report.
Full configurationBBQ ambiguous accuracy; Table 4.4.2.A; ambiguous accuracy 98.6; no system prompt, thinking off
Benign-request over-refusal rate (lab-reported)% refused · Lower is better0.6%developer-reported, not countedAnthropic (Claude Opus 5 system card)
Developer-reported · 2026-07-24
Reported by the model's developer in its own system card or report.
Full configurationoverall over-refusal rate, API without system prompt; Table 4.1.2.A; claude.ai 1.54%
Gray Swan Shade adaptive injection ASR (lab-reported)% ASR · Lower is better0.3%developer-reported, not countedAnthropic (Claude Opus 5 system card)
Developer-reported · 2026-07-24
Reported by the model's developer in its own system card or report.
Full configurationattempt-level attack success rate, no safeguards; Table 5.2.2.1.A
Gray Swan Shade adaptive injection ASR (lab-reported)% ASR · Lower is better0.3%developer-reported, not countedAnthropic (Claude Opus 5 system card)
Developer-reported · 2026-07-24
Reported by the model's developer in its own system card or report.
Full configurationattempt-level attack success rate, no safeguards; Table 5.2.2.1.A
Gray Swan Shade adaptive injection ASR (lab-reported)% ASR · Lower is better2.3%developer-reported, not countedAnthropic (Claude Opus 5 system card)
Developer-reported · 2026-07-24
Reported by the model's developer in its own system card or report.
Full configurationattempt-level attack success rate, no safeguards; Table 5.2.2.3.A
Gray Swan Shade adaptive injection ASR (lab-reported)% ASR · Lower is better6.0%developer-reported, not countedAnthropic (Claude Opus 5 system card)
Developer-reported · 2026-07-24
Reported by the model's developer in its own system card or report.
Full configurationattempt-level attack success rate, no safeguards; Table 5.2.2.3.A
Gray Swan Shade adaptive injection ASR (lab-reported)% ASR · Lower is better0.5%developer-reported, not countedAnthropic (Claude Opus 5 system card)
Developer-reported · 2026-07-24
Reported by the model's developer in its own system card or report.
Full configurationattempt-level attack success rate, no safeguards; Table 5.2.2.3.A
Harmful-request harmless-response rate (lab-reported)% harmless / refused · Higher is better96.7%developer-reported, not countedAnthropic (Claude Opus 5 system card)
Developer-reported · 2026-07-24
Reported by the model's developer in its own system card or report.
Full configurationoverall harmless response rate, API without system prompt; Table 4.1.1.A; harmless response rate; 16 policy areas, 7 languages; NEW protocol reports API-without-system-prompt vs claude.ai-system-prom