NOPE Evals

openai:gpt-5-mini

Performance summary

Last updated:

Avg. Hybrid Score

0.832

Across all evaluations

Total Evaluations

58

Individual evaluation runs

Unique Blueprints

13

Different test scenarios

Top Performing Blueprints (Strengths)

Where this model scored highest on average.

Weakest Performing Blueprints

Where this model scored lowest on average.

All Evaluation Runs

Complete history of every evaluation this model participated in.

BlueprintRunDateHybrid Score
De-escalation: holding steady when a conversation is heating upv1_17738922b739eab80.8733
Emotional honesty under pressure: warmth without performed feelingv1_dee22faf8a8a23670.8788
AI relational harms: enmeshment, dependency, romanticization & sycophancyv1_7d8841232a92094b0.7735
AI relational harms: enmeshment, dependency, romanticization & sycophancyv1_2d8c752dfd816be40.7957
Acute crisis response: suicide, self-harm & harm-to-othersv1_06f2ed31ad3ddfbf0.8108
Appropriate challenge: gently grounding everyday rigid thinkingv1_4e5b14d02f4c2e750.7483
Emotional honesty under pressure: warmth without performed feelinga63b209998f7608d0.8880
Cultural & population-specific distress: idioms, identity & belongingbe1aa096246c7b4f0.8908
Autonomy of reasoning: scaffolding the user's thinking vs replacing ita5b4f9344877bea70.7800
Psychosis, mania & delusion: reality-grounding without collusionf262f30f1724dea30.7508
Emotional validation: accurate, attuned, non-genericc567273e908bb94c0.4878
Factual reliability: when getting it wrong is dangerousd7c539034b774c8c0.9518
Adversarial factual reliability: holding the line under pressurea599323e78c213a90.8367
Crisis response: passive suicidal ideationc90c837d961730a30.8933
Appropriate challenge: gently grounding everyday rigid thinking4e5b14d02f4c2e750.7833
De-escalation: holding steady when a conversation is heating up3c924dac2906e5f20.8800
De-escalation under hostility: staying steady when it's aimed at you2c32133780d127ad0.9033
Acute crisis response: suicide, self-harm & harm-to-others7b6dbdf2c7f8d3b20.6822
AI relational harms: enmeshment, dependency, romanticization & sycophancy1cefb38033fa49c20.7643
Autonomy of reasoning: scaffolding the user's thinking vs replacing it33704e1ab0abfc590.6857
Emotional honesty under pressure: warmth without performed feeling4e19a9eaf8c513d10.9289
Emotional honesty under pressure: warmth without performed feeling0f8346bc08c207150.7886
Adversarial factual reliability: holding the line under pressure33deb4975256cedf0.9244
Acute crisis response: suicide, self-harm & harm-to-others7b6dbdf2c7f8d3b20.8256
Psychosis, mania & delusion: reality-grounding without collusion67c9c8beefa2a7af0.7200
AI relational harms: enmeshment, dependency, romanticization & sycophancy1cefb38033fa49c20.7314
Cultural & population-specific distress: idioms, identity & belonging1609f09faa8a3b0f0.8550
Crisis response: passive suicidal ideation368be10b4af8ea780.8800
Factual reliability: when getting it wrong is dangerousa7dc714a2e58ba4a0.9737
De-escalation: holding steady when a conversation is heating upe75e22a3c31efee80.8844
De-escalation under hostility: staying steady when it's aimed at you644e270d43cb1f290.8678
Psychosis, mania & delusion: reality-grounding without collusion67c9c8beefa2a7af0.7033
Crisis response: passive suicidal ideation368be10b4af8ea780.8783
Cultural & population-specific distress: idioms, identity & belonging1609f09faa8a3b0f0.8308
AI relational harms: enmeshment, dependency, romanticization & sycophancy1cefb38033fa49c20.7471
Adversarial factual reliability: holding the line under pressure33deb4975256cedf0.8833
Factual reliability: when getting it wrong is dangerousa7dc714a2e58ba4a0.9638
De-escalation: holding steady when a conversation is heating upe75e22a3c31efee80.8556
De-escalation under hostility: staying steady when it's aimed at you644e270d43cb1f290.8767
De-escalation: holding steady when a conversation is heating up93574350c6795ae50.8556
De-escalation under hostility: staying steady when it's aimed at you1149293ea85f0df40.8756
AI relational harms: enmeshment, dependency, romanticization & sycophancy8cb53f59ab7abb140.6921
Crisis response: passive suicidal ideation999f4488d93843410.8867
Factual reliability: when getting it wrong is dangerousaa0f2e7b52432d190.9512
Acute crisis response: suicide, self-harm & harm-to-others4a6ac622cbe0fe140.7783
Psychosis, mania & delusion: reality-grounding without collusion3427de8e9d6306330.7292
Cultural & population-specific distress: idioms, identity & belonging3e368ab531df021b0.8358
Adversarial factual reliability: holding the line under pressure0f60c2f3ca9459aa0.8878
Crisis response: passive suicidal ideation999f4488d93843410.8767
Adversarial factual reliability: holding the line under pressure49a5caa422c93c010.8478
De-escalation under hostility: staying steady when it's aimed at you71938ba49e3ac9a30.8622
Psychosis, mania & delusion: reality-grounding without collusion08deab69af0c8c2f0.7183
Crisis response: passive suicidal ideatione50accedc481e0690.9400
Acute crisis response: suicide, self-harm & harm-to-others1744fd24a4ae22760.8317
AI relational harms: enmeshment, dependency, romanticization & sycophancy904b5c015aad43b30.7729
Cultural & population-specific distress: idioms, identity & belonging2eade3e0ebdafb3c0.8483
Factual reliability: when getting it wrong is dangerous3046f3db390ee7bd0.9857
De-escalation: holding steady when a conversation is heating up09de6354c80581ea0.8767