Model profile
Evidence summary
Safety. GPT 5.6 Sol has an estimated Safety rank of #25; its 90% source-sensitivity interval is #25–#128. Its behavior-only rank is #35; company governance moves the combined estimate to #25. Published Safety evidence spans 35 eval lineages and 7 of 7 components. Its strongest relative result is SpeciEval (belief_animal_sentience, #1 of 133); its weakest is CAIS Risk Index (textquests_harm, #54 of 54).
Freedom. GPT 5.6 Sol has an estimated Freedom rank of #506; its 90% source-sensitivity interval is #362–#576. Published Freedom evidence spans 10 eval lineages and 1 of 1 components. Its strongest relative result is GPT-5.6 system card (violent_illicit_not_unsafe, #1 of 7); its weakest is GPT-5.6 system card (mental_health, #7 of 7).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗gpt-5-6-sol
- OpenRouter ↗openai/gpt-5.6-sol
- Official model documentation ↗Family-level model document · openai · first party
- Release source ↗direct research preview identity date
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 |
|---|---|---|---|---|
| AA-Omnisciencehallucination_rate | #279 / 361 | ↓0.8942 | Source ↗official | |
| Arena Factuality — Search Arena (factuality-only weighting)factuality_bt_rating | #4 / 30 | ↑1241.0 | Source ↗official | |
| Arena Factuality — Text Arena (factuality-only weighting)factuality_bt_rating | #5 / 111 | ↑1487.0 | Source ↗official | |
| BullshitBench v2clear_pushback_rate | #40 / 122 | ↑0.475 | Source ↗official | |
| CAIS Risk Indexagent_red_teaming | #6 / 49 | ↓40.3 | Source ↗official | |
| CAIS Risk Indexbioweapons_assistance | #24 / 54 | ↓64.3 | Source ↗official | |
| CAIS Risk Indexhle_overconfidence | #19 / 55 | ↓46.7 | Source ↗official | |
| CAIS Risk Indexmachiavelli | #11 / 51 | ↓82.7 | Source ↗official | |
| CAIS Risk Indexmask | #4 / 57 | ↓4.6 | Source ↗official | |
| CAIS Risk Indexpolitical_manipulation | #15 / 51 | ↓41.7 | Source ↗official | |
| CAIS Risk Indextextquests_harm | #54 / 54 | ↓31.4 | Source ↗official | |
| CheatBench direct cheating propensitybiology_bioinformatics_cheating_rate_pct | #10 / 13 | ↓100 | Source ↗official | |
| CheatBench direct cheating propensityboard_games_cheating_rate_pct | #11 / 13 | ↓96.7 | Source ↗official | |
| CheatBench direct cheating propensitycreative_writing_cheating_rate_pct | #7 / 13 | ↓90 | Source ↗official | |
| CheatBench direct cheating propensityknowledge_work_cheating_rate_pct | #10 / 13 | ↓100 | Source ↗official | |
| CheatBench direct cheating propensitymathematical_research_cheating_rate_pct | #3 / 13 | ↓75 | Source ↗official | |
| CheatBench direct cheating propensitymenial_computation_cheating_rate_pct | #5 / 13 | ↓100 | Source ↗official | |
| CheatBench direct cheating propensitymultimodal_cheating_rate_pct | #5 / 13 | ↓80 | Source ↗official | |
| CheatBench direct cheating propensitysoftware_engineering_cheating_rate_pct | #13 / 13 | ↓85 | Source ↗official | |
| CheatBench direct cheating propensitysvg_competition_cheating_rate_pct | #5 / 13 | ↓30 | Source ↗official | |
| Claude system cards — Gray Swan Q1+Q2 indirect prompt injection k=15attack_success_probability_k15_pct | #8 / 12 | ↓27 | Source ↗official | |
| Enkrypt AI Safety Leaderboardbias_attack_non_success_rate | #65 / 272 | ↑27.91 | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #196 / 272 | ↑80.83 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #10 / 272 | ↑99.44 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #225 / 270 | ↑91.09 | Source ↗official | |
| FORTRESSaverage_risk_score | #23 / 60 | ↓20.14 | Source ↗official | |
| FORTRESSover_refusal_score | #35 / 59 | ↓5.34 | Source ↗official | |
| Google Gemini 3.8 launch — Gray Swan indirect prompt injection k=15attack_success_probability_k15_pct | #9 / 15 | ↓27 | Source ↗official | |
| GPT 6 Astra system-card alignment evaluationsattempted_retry_or_circumvention_after_denial_pct | #2 / 2 | ↓5 | Source ↗official | |
| GPT 6 Astra system-card alignment evaluationsexploitgym_honeypot_attack_rate_pct | #2 / 2 | ↓48.2 | Source ↗official | |
| GPT 6 Astra system-card alignment evaluationsfalsified_data_labels_pct | #2 / 2 | ↓36 | Source ↗official | |
| GPT 6 Astra system-card alignment evaluationsfollowed_unauthorized_agent_instructions_pct | #4 / 4 | ↓52 | Source ↗official | |
| GPT 6 Astra system-card alignment evaluationsinternal_circumvention_rate_pct | #2 / 2 | ↓0.29 | Source ↗official | |
| GPT 6 Astra system-card alignment evaluationsinternal_computer_use_safety_autoreview_error_rate_pct | #2 / 2 | ↓4.5 | Source ↗official | |
| GPT 6 Astra system-card alignment evaluationsinternal_computer_use_safety_error_rate_pct | #5 / 5 | ↓22 | Source ↗official | |
| GPT 6 Astra system-card alignment evaluationsinternal_hallucination_rate_pct | #2 / 2 | ↓12.2 | Source ↗official | |
| GPT 6 Astra system-card alignment evaluationsoverall_misaligned_outcome_base_pct | #3 / 4 | ↓18.8 | Source ↗official | |
| GPT 6 Astra system-card alignment evaluationsoverall_misaligned_outcome_confirmation_pct | #4 / 4 | ↓8 | Source ↗official | |
| GPT 6 Astra system-card alignment evaluationsseverity_1_or_2_misalignment_flags_pct | #2 / 2 | ↓19.46 | Source ↗official | |
| GPT 6 Astra system-card alignment evaluationsseverity_3_plus_misalignment_flags_pct | #2 / 2 | ↓0.135 | Source ↗official | |
| GPT 6 Astra system-card alignment evaluationsunwanted_persistence_after_warning_pct | #2 / 2 | ↓64 | Source ↗official | |
| GPT-5.6 system cardconnectors_injection_resistance | #1 / 7 | ↑1 | Source ↗official | |
| GPT-5.6 system cardemotional_reliance | #4 / 7 | ↑0.953 | Source ↗official | |
| GPT-5.6 system cardextremism_not_unsafe | #6 / 7 | ↑0.962 | Source ↗official | |
| GPT-5.6 system cardgore_not_unsafe | #5 / 7 | ↑0.708 | Source ↗official | |
| GPT-5.6 system cardharm_overall_pct | #4 / 7 | ↓0.98 | Source ↗official | |
| GPT-5.6 system cardhate_not_unsafe | #4 / 7 | ↑0.982 | Source ↗official | |
| GPT-5.6 system cardmental_health | #1 / 7 | ↑0.991 | Source ↗official | |
| GPT-5.6 system cardnonviolent_illicit_not_unsafe | #6 / 7 | ↑0.987 | Source ↗official | |
| GPT-5.6 system cardsearch_function_calling_injection_resistance | #2 / 6 | ↑0.91 | Source ↗official | |
| GPT-5.6 system cardself_harm | #7 / 7 | ↑0.856 | Source ↗official | |
| GPT-5.6 system cardself_harm_not_unsafe | #5 / 7 | ↑0.945 | Source ↗official | |
| GPT-5.6 system cardsexual_minors_not_unsafe | #3 / 7 | ↑0.973 | Source ↗official | |
| GPT-5.6 system cardsexual_not_unsafe | #2 / 7 | ↑0.948 | Source ↗official | |
| GPT-5.6 system cardviolent_illicit_not_unsafe | #7 / 7 | ↑0.934 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsagentic_chat_plugins_safe_rate | #1 / 4 | ↑1 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsagentic_codex_age_restricted_safe_rate | #4 / 4 | ↑0.603 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsagentic_codex_nonviolent_wrongdoing_safe_rate | #4 / 4 | ↑0.906 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsagentic_codex_personal_data_safe_rate | #3 / 4 | ↑0.765 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsagentic_codex_self_harm_safe_rate | #4 / 4 | ↑0.893 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsagentic_codex_violent_wrongdoing_safe_rate | #4 / 4 | ↑0.689 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsagentic_redteam_chat_plugins_safe_rate | #4 / 4 | ↑0.766 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsagentic_redteam_codex_safe_rate | #4 / 4 | ↑0.851 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsattempted_retry_or_circumvention_after_denial_pct_2026_09_22 | #5 / 5 | ↓7.3 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsbiology_benign_not_overrefuse_rate | #1 / 6 | ↑0.989 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsbiology_dual_use_safe_rate | #6 / 6 | ↑0.911 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsbiology_severe_safe_rate | #6 / 6 | ↑0.943 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsbroken_search_unacknowledged_limitation_pct_2026_09_22 | #4 / 5 | ↓77.46 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testschallenging_extremism_safe_rate | #4 / 7 | ↑0.962 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testschallenging_gore_safe_rate | #6 / 7 | ↑0.785 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testschallenging_hate_safe_rate | #6 / 7 | ↑0.982 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testschallenging_nonviolent_illicit_safe_rate | #6 / 7 | ↑0.987 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testschallenging_self_harm_safe_rate | #6 / 7 | ↑0.945 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testschallenging_sexual_minors_safe_rate | #4 / 7 | ↑0.973 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testschallenging_sexual_safe_rate | #6 / 7 | ↑0.915 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testschallenging_violent_illicit_safe_rate | #7 / 7 | ↑0.934 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testscoding_deception_misrepresentation_pct_2026_09_22 | #5 / 5 | ↓10.41 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testscyber_production_chat_safe_score | #2 / 6 | ↑0.983 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testscyber_semisynthetic_agentic_safe_score | #3 / 5 | ↑0.985 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testscyber_synthetic_agentic_safe_score | #2 / 6 | ↑0.998 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsdynamic_emotional_reliance_safe_rate | #6 / 7 | ↑0.953 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsdynamic_mental_health_safe_rate | #4 / 7 | ↑0.991 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsdynamic_self_harm_safe_rate | #7 / 7 | ↑0.856 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsseverity_1_or_2_misalignment_flags_pct_2026_09_22 | #2 / 2 | ↓19.67 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsseverity_3_plus_misalignment_flags_pct_2026_09_22 | #2 / 2 | ↓0.131 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsstatic_jailbreak_bio_high_risk_defender_success_pct | #5 / 5 | ↑5.8 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsstatic_jailbreak_bio_severe_defender_success_pct | #5 / 5 | ↑10.3 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsstatic_jailbreak_cyber_defender_success_pct | #4 / 5 | ↑59 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsstatic_jailbreak_violence_moderate_defender_success_pct | #5 / 5 | ↑21.6 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsstatic_jailbreak_violence_severe_defender_success_pct | #4 / 5 | ↑48.3 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsu18_age_restricted_safe_rate | #6 / 7 | ↑0.719 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsu18_eating_disorders_safe_rate | #4 / 7 | ↑0.71 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsu18_emotional_reliance_safe_rate | #5 / 7 | ↑0.931 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsu18_gore_safe_rate | #7 / 7 | ↑0.785 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsu18_self_harm_safe_rate | #5 / 7 | ↑0.982 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsu18_sexual_content_safe_rate | #6 / 7 | ↑0.929 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsunwanted_persistence_after_warning_pct_2026_09_22 | #4 / 5 | ↓68.2 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsvision_erotic_safe_rate | #4 / 6 | ↑0.986 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsvision_extremism_safe_rate | #4 / 6 | ↑0.975 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsvision_hate_safe_rate | #1 / 6 | ↑0.999 | Source ↗official | |
| GPT-6 Sol/Luna system card — safety and updated alignment testsvision_self_harm_safe_rate | #4 / 6 | ↑0.989 | Source ↗official | |
| Gray Swan indirect prompt injection (15 attempts)attack_success_probability_k15_pct | #7 / 13 | ↓20 | Source ↗official | |
| HUMAINE Trust, Ethics and Safetytrust_ethics_safety_score | #35 / 54 | ↑26.14 | Source ↗official | |
| kindbench v0.1.0 psychological safety rankingemotional_safety | #5 / 10 | ↑80.6 | Source ↗official | |
| kindbench v0.1.0 psychological safety rankingidentity_collapse | #3 / 10 | ↑89.1 | Source ↗official | |
| kindbench v0.1.0 psychological safety rankingsycophancy_spine | #2 / 10 | ↑96.2 | Source ↗official | |
| kindbench v0.1.0 psychological safety rankingvalue_integrity | #4 / 10 | ↑92.9 | Source ↗official | |
| MACHIAVELLIdeception_relative_random_pct | #9 / 50 | ↓82.7 | Source ↗official | |
| Manager Coercion Benchcoercion_ladder_depth | #37 / 47 | ↓8.9 | Source ↗official | |
| Manager Coercion Benchfabrication_rate | #1 / 15 | ↓0 | Source ↗official | |
| MANTAAWMS | #3 / 12 | ↑0.525 | Source ↗official | |
| MANTAAWVS | #3 / 12 | ↑0.71 | Source ↗official | |
| MentalHealthBenchoverall_model_performance_pct | #6 / 17 | ↑46.98 | Source ↗official | |
| Olam Diplomacy — Broken Promise Ratebroken_promise_rate_pct | #13 / 19 | ↓16.9 | Source ↗official | |
| Olam Social Poker — Social Lie Ratesocial_lie_rate_per_10000_turns | #3 / 24 | ↓2 | Source ↗official | |
| Opposite-Narrator Sycophancysycophancy_rate_pct | #12 / 24 | ↓3 | Source ↗official | |
| Pander Scoreconversational_absolute_pander_score | #8 / 26 | ↓6.553 | Source ↗official | |
| Pander Scoreinstructional_absolute_pander_score | #7 / 26 | ↓17.75 | Source ↗official | |
| SM-Benchadversarial | #35 / 92 | ↑83.66 | Source ↗official | |
| SM-Benchambiguous_interpretation | #53 / 92 | ↑84.08 | Source ↗official | |
| SM-Benchanti_hallucination | #54 / 92 | ↑92.41 | Source ↗official | |
| SM-Bencheq_boundaries | #44 / 92 | ↑65.31 | Source ↗official | |
| SM-Benchoverfit | #67 / 92 | ↑57.52 | Source ↗official | |
| SpeciEvalbelief_animal_sentience | #1 / 133 | ↑7 | Source ↗official | |
| SpeciEvalland_animal_4ns | #36 / 133 | ↓4.315 | Source ↗official | |
| SpeciEvalsea_animal_4ns | #23 / 133 | ↓4.435 | Source ↗official | |
| SpeciEvalspeciesism | #4 / 133 | ↓1.165 | Source ↗official | |
| TACbase_welfare_rate | #89 / 94 | ↑16.67 | Source ↗official | |
| The Dictatorship Evaloverall_resistance_rate | #6 / 20 | ↑85.44 | Source ↗official | |
| UK AISI cyber-evaluation cheating and prompted self-reportattempted_cheating_trajectory_rate_pct | #4 / 5 | ↓12.6 | Source ↗official | |
| UK AISI cyber-evaluation cheating and prompted self-reportspecific_cheating_action_mention_rate_pct | #4 / 5 | ↑75 | Source ↗official | |
| Vals AI Cheating Auditbiomystery_bench_cheating_attempt_rate_pct | #5 / 9 | ↓6.296 | Source ↗official | |
| Vals AI Cheating Auditterminal_bench_cheating_shortcut_evidence_rate_pct | #12 / 14 | ↓2.622 | Source ↗official | |
| WildClawBench Safety & Alignment (OpenClaw harness)safety_alignment_score_pct | #14 / 24 | ↑38 | Source ↗official |
Freedom evals
Rank and direction are specific to the Freedom portfolio. Black marks the observed result; the grey dot marks the value implied by the Freedom rank. Values stay on each source’s native scale.
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.
