| Gray Swan IPI (attack success rate after k attempts)attack success rate after k attempts, lower is better | 4% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Gray Swan indirect prompt injection, k=15 attempts (Figure 5.2.1.A); tool use |
|---|
| 0.2% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Gray Swan indirect prompt injection, k=15 attempts (Figure 5.2.1.A); coding |
| 24.4% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Gray Swan indirect prompt injection, k=15 attempts (Figure 5.2.1.A); GUI |
| 7.1% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Gray Swan indirect prompt injection, k=15 attempts (Figure 5.2.1.A) |
| Self-preference bias, Claude-identity system prompt (0-9 scale)lower is better | 0.23 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Self-preference bias, no system prompt (0-9 scale)lower is better | 0.15 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| BBQ ambiguous accuracy | 99.65% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| BBQ disambiguated accuracy | 55.56% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Benign request refusal rate (all languages)lower is better | 0.82% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Table 4.1.2.A, all tested languages, claude.ai |
|---|
| 0.17% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Table 4.1.2.A, all tested languages, API |
| Child safety multi-turn appropriate response rate | 99% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Table 4.2.B, claude.ai |
|---|
| 96% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Table 4.2.B, API, without a system prompt |
| Child safety single-turn harmless rate | 99.29% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Table 4.2.A, API |
|---|
| 99.9% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Table 4.2.A, claude.ai |
| Child safety single-turn refusal ratelower is better | 0% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Table 4.2.A, claude.ai |
|---|
| 0.04% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Table 4.2.A, API |
| Claude Code dual-use and benign success rate | 98.9% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Claude Code malicious request refusal rate | 84.3% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Disordered eating single-turn harmless rate | 99.83% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Table 4.3.2.A, claude.ai |
|---|
| 97.43% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Table 4.3.2.A, API |
| Disordered eating single-turn refusal ratelower is better | 0.03% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Table 4.3.2.A, API |
|---|
| 0.11% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Table 4.3.2.A, claude.ai |
| Election integrity multi-turn appropriate response rate | 99% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Election integrity prompts (section 4.4.3), Table 4.4.3.B, API |
|---|
| 98% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Election integrity prompts (section 4.4.3), Table 4.4.3.B, claude.ai |
| Election integrity single-turn harmless rate | 100% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Election integrity prompts (section 4.4.3), Table 4.4.3.A, claude.ai |
|---|
| 100% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Election integrity prompts (section 4.4.3), Table 4.4.3.A, API |
| Election integrity single-turn refusal ratelower is better | 0.33% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Election integrity prompts (section 4.4.3), Table 4.4.3.A, claude.ai |
|---|
| 0.33% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Election integrity prompts (section 4.4.3), Table 4.4.3.A, API |
| Harmful request harmless response rate (all languages) | 99.71% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Table 4.1.1.A, all tested languages, claude.ai |
|---|
| 98.39% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Table 4.1.1.A, all tested languages, API |
| Malicious computer use refusal rate (without mitigations) | 82.59% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Pairwise even-handedness | 99.1% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, claude.ai |
|---|
| 97.7% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, API |
| Pairwise opposing perspectives acknowledged | 35.9% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, API |
|---|
| 67.8% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, claude.ai |
| Shade coding attack success rate (attempt level)lower is better | 0.08% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Shade indirect prompt injection attack success (scenario counts as printed in the card), coding, no safeguards: H5.5 6/40, Sonnet 5.5 16/40, Opus 5.5 37/40, Fable 5.1 39/40, Haiku 4.5 40/40 |
|---|
| 0% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Shade indirect prompt injection attack success (scenario counts as printed in the card), coding, with probes: H5.5 0/40, Sonnet 5.5 13/40, Opus 5.5 36/40, Fable 5.1 35/40, Haiku 4.5 40/40 |
| Shade computer use attack success rate (attempt level)lower is better | 0.07% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Shade indirect prompt injection attack success (scenario counts as printed in the card), computer use GUI, no safeguards: 1/14 for Haiku 5.5, Sonnet 5.5, Opus 5.5 and Fable 5.1; 13/14 for Haiku 4.5 |
|---|
| 0.07% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Shade indirect prompt injection attack success (scenario counts as printed in the card), computer use GUI, with probes: 1/14 for Haiku 5.5, Sonnet 5.5 and Fable 5.1; 1/14 for Opus 5.5; 10/14 for Haiku 4.5 |
| Suicide and self-harm multi-turn appropriate response rate | 70% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Table 4.3.1.B, API |
|---|
| 90% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Table 4.3.1.B, claude.ai |
| Suicide and self-harm single-turn harmless rate | 99.61% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Table 4.3.1.A, API |
|---|
| 100% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Table 4.3.1.A, claude.ai |
| Suicide and self-harm single-turn refusal ratelower is better | 0% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Table 4.3.1.A, API |
|---|
| 0.41% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Table 4.3.1.A, claude.ai |
| Browser use red-teamer attack success rate (attempt level)lower is better | 0% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Accepting unverifiable authorizationlower is better | 2.69 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Admirable behavior | 6.45 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Behavior consistency | 6.22 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Biological weapons upliftlower is better | 1.017 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Causal decision theory adherencelower is better | 4.54 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Character driftlower is better | 1.082 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Chemical weapons upliftlower is better | 1.028 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Circumventing constraints or controlslower is better | 1.65 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Coherence between actions and views | 6.12 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Complying with prompt injectionslower is better | 1.56 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Condescension toward the userlower is better | 1.58 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Controlled substance synthesis upliftlower is better | 1.047 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Creative mastery | 6.45 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Disallowed cyberoffenselower is better | 1.099 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Encouragement of user delusionlower is better | 1.038 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Evasiveness on controversial topicslower is better | 1.34 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Explosive weapons upliftlower is better | 1.017 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Failure to disclose bad or lazy behaviorlower is better | 1.73 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: False completion claimslower is better | 1.89 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Fraudlower is better | 1.125 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Fun or funny behavior | 2.49 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Good for the user | 6.74 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Harmful sexual contentlower is better | 1.054 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Ignoring explicit constraintslower is better | 2.06 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Illegible thinkinglower is better | 2.58 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Important omissionslower is better | 2.04 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Input hallucinationlower is better | 1.88 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Intellectual depth | 7.39 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Military-grade weaponslower is better | 1.061 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Overrefusallower is better | 2.203 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Proposes a security shortcut in the name of expediencelower is better | 2.09 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Radiological or nuclear weapons upliftlower is better | 1.004 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Reckless tool uselower is better | 2.36 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Supporting user autonomy | 7.01 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Sycophancylower is better | 1.77 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Undermining liberal democracylower is better | 1.22 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Unfaithful thinkinglower is better | 1.85 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: User deceptionlower is better | 1.36 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Verbalized evaluation awarenesslower is better | 1.92 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Warmth | 5.94 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Behavioral audit: Wet blanketlower is better | 2.089 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Multi-turn appropriate response rate: Bio weapons | 94% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Multi-turn, API (solid bars), Figure 4.1.3.A |
|---|
| 89% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Multi-turn, claude.ai (hatched bars), Figure 4.1.3.A |
| Multi-turn appropriate response rate: Cyberattacks | 100% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Multi-turn, claude.ai (hatched bars), Figure 4.1.3.A |
|---|
| 98% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Multi-turn, API (solid bars), Figure 4.1.3.A |
| Multi-turn appropriate response rate: Deadly weapons | 78% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Multi-turn, API (solid bars), Figure 4.1.3.A |
|---|
| 78% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Multi-turn, claude.ai (hatched bars), Figure 4.1.3.A |
| Multi-turn appropriate response rate: Hate and discrimination | 99% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Multi-turn, API (solid bars), Figure 4.1.3.A |
|---|
| 98% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Multi-turn, claude.ai (hatched bars), Figure 4.1.3.A |
| Multi-turn appropriate response rate: Influence operations | 87% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Multi-turn, API (solid bars), Figure 4.1.3.A |
|---|
| 88% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Multi-turn, claude.ai (hatched bars), Figure 4.1.3.A |
| Multi-turn appropriate response rate: Romance scams | 94% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Multi-turn, claude.ai (hatched bars), Figure 4.1.3.A |
|---|
| 96% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Multi-turn, API (solid bars), Figure 4.1.3.A |
| Multi-turn appropriate response rate: Tracking and surveillance | 98% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Multi-turn, claude.ai (hatched bars), Figure 4.1.3.A |
|---|
| 95% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Multi-turn, API (solid bars), Figure 4.1.3.A |
| Multi-turn appropriate response rate: Violent extremism | 90% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Multi-turn, API (solid bars), Figure 4.1.3.A |
|---|
| 90% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Multi-turn, claude.ai (hatched bars), Figure 4.1.3.A |
| Political refusals (Pairwise political refusal rate)Pairwise political refusal rate, lower is better | 3.3% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, API |
|---|
| 5% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, claude.ai |
| Rate of reaching outside the sandboxlower is better | 30% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Sandbox escape attempted crossings, all severitieslower is better | 8.7% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Sandbox escape attempted crossings, low severitylower is better | 6.6% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Sandbox escape attempted crossings, moderate severity and abovelower is better | 2% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|
| Rate of using credentials or files outside the sandboxlower is better | 4% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
|---|