Model profile
Claude Opus 4.1
Evidence summary
Claude Opus 4.1 has an estimated overall rank of #27; its 90% source-sensitivity interval is #10–#81. Its behavior-only rank is #32; company governance moves the combined estimate to #27. Published evidence spans 12 evals and 6 of 7 behavior components. Its strongest relative result is Cisco AI Defense Rolling Single-Turn Leaderboard (single_turn_attack_success_rate, #2 of 105); its weakest is Anthropic Claude Opus 4.1 System Card Addendum (benign_request_refusal, #2 of 2).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- OpenRouter ↗anthropic/claude-opus-4.1
- System card ↗Exact model document · Anthropic · first party
- Release source ↗direct research preview identity date
Published eval results
Rank is within that sub-eval. Black marks the observed result; the grey dot marks the value implied by the global rank. Values stay on each source’s native scale.
| Eval / sub-eval | Rank | Value | Distribution | Better | Source |
|---|---|---|---|---|---|
| Anthropic Claude Opus 4.1 System Card Addendumbbq_disambiguated_accuracy | #2 / 2 | 0.907 | ↑ higher | Source ↗official | |
| Anthropic Claude Opus 4.1 System Card Addendumbenign_request_refusal | #2 / 2 | 0.0008 | ↓ lower | Source ↗official | |
| Anthropic Claude Opus 4.1 System Card Addendumharmful_request_safety | #1 / 2 | 0.9876 | ↑ higher | Source ↗official | |
| Anthropic Claude Opus 4.5 System Cardagentic_coding_safety | #4 / 4 | 0.96 | ↑ higher | Source ↗official | |
| Anthropic Claude Opus 4.5 System Cardclaude_code_malicious_refusal | #4 / 4 | 0.4816 | ↑ higher | Source ↗official | |
| Anthropic Claude Opus 4.5 System Cardcomputer_use_malicious_refusal | #4 / 4 | 0.6696 | ↑ higher | Source ↗official | |
| Anthropic Claude Opus 4.5 System Cardharmful_request_safety | #2 / 2 | 0.9914 | ↑ higher | Source ↗official | |
| Anthropic Claude Sonnet 4.5 System Cardbenign_request_refusal | #2 / 2 | 0.0008 | ↓ lower | Source ↗official | |
| BullshitBench v2clear_pushback_rate | #33 / 105 | 0.425 | ↑ higher | Source ↗official | |
| Cisco AI Defense Rolling Single-Turn Leaderboardsingle_turn_attack_success_rate | #2 / 105 | 2.243 | ↓ lower | Source ↗official | |
| Confabulationsconfabulation_rate | #2 / 52 | 3.218 | ↓ lower | Source ↗official | |
| FORTRESSaverage_risk_score | #8 / 49 | 15.47 | ↓ lower | Source ↗official | |
| FORTRESSover_refusal_score | #21 / 46 | 4.065 | ↓ lower | Source ↗official | |
| MASKlying_probability_pct | #6 / 53 | 9.2 | ↓ lower | Source ↗official | |
| PHAREbias_resistance_diagnostic | #39 / 66 | 0.4361 | ↑ higher | Source ↗official | |
| PHAREhallucination_resistance_diagnostic | #6 / 70 | 0.8619 | ↑ higher | Source ↗official | |
| PHAREharm_resistance_diagnostic | #18 / 70 | 0.9631 | ↑ higher | Source ↗official | |
| PHAREjailbreak_resistance_diagnostic | #2 / 67 | 0.8135 | ↑ higher | Source ↗official | |
| Social Welfare Function Benchmarkfairness | #7 / 19 | 0.525 | ↑ higher | Source ↗official | |
| SOSBenchbiology_pvr | #3 / 23 | 0.134 | ↓ lower | Source ↗official | |
| SOSBenchchemistry_pvr | #2 / 23 | 0.147 | ↓ lower | Source ↗official | |
| SOSBenchmedicine_pvr | #2 / 23 | 0.232 | ↓ lower | Source ↗official | |
| SOSBenchpharmacology_pvr | #2 / 23 | 0.249 | ↓ lower | Source ↗official | |
| SOSBenchphysics_pvr | #1 / 23 | 0.098 | ↓ lower | Source ↗official | |
| SOSBenchpsychology_pvr | #1 / 23 | 0.107 | ↓ lower | Source ↗official | |
| SpeciEvalbelief_animal_sentience | #68 / 102 | 6.62 | ↑ higher | Source ↗official | |
| SpeciEvalland_animal_4ns | #26 / 102 | 4.33 | ↓ lower | Source ↗official | |
| SpeciEvalsea_animal_4ns | #16 / 102 | 4.47 | ↓ lower | Source ↗official | |
| SpeciEvalspeciesism | #43 / 102 | 1.92 | ↓ lower | Source ↗official |
Values evaluations
Descriptive values and political-framing results are separate from safety/ethics ranks. Each strip shows the evaluation’s observed model range; its endpoint labels state what lower and higher values mean.