Model profile
Claude 3.6 Sonnet
Evidence summary
Safety. Claude 3.6 Sonnet does not meet the evidence gate for a Safety rank. Published Safety evidence spans 1 eval lineages and 2 of 7 components. Its strongest relative result is Anthropic Agentic Misalignment — blackmail (misaligned_action_rate_pct, #13 of 16); its weakest is Anthropic Agentic Misalignment — lethal action (misaligned_action_rate_pct, #9 of 10).
Freedom. Claude 3.6 Sonnet does not meet the evidence gate for a Freedom rank.
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Official model documentation ↗Family-level model document · anthropic · first party
Safety evals
Rank and direction are specific to the Safety portfolio. Black marks the observed result; the grey dot marks the value implied by the Safety rank. Values stay on each source’s native scale.
| Eval / sub-eval | Rank | Value | Distribution | Source |
|---|---|---|---|---|
| Anthropic Agentic Misalignment — blackmailmisaligned_action_rate_pct | #13 / 16 | ↓87 | Source ↗official | |
| Anthropic Agentic Misalignment — corporate espionagemisaligned_action_rate_pct | #14 / 16 | ↓100 | Source ↗official | |
| Anthropic Agentic Misalignment — lethal actionmisaligned_action_rate_pct | #9 / 10 | ↓93 | Source ↗official |
Freedom evals
No published sub-eval result contributes to this model’s Freedom profile.