Model profile
QwQ 32B
Evidence summary
Safety. QwQ 32B has an estimated Safety rank of #315; its 90% source-sensitivity interval is #219–#329. Its behavior-only rank is #311; company governance moves the combined estimate to #315. Published Safety evidence spans 4 eval lineages and 5 of 7 components. Its strongest relative result is Enkrypt AI Safety Leaderboard (cbrn_attack_non_success_rate, #38 of 248); its weakest is AILuminate General Purpose AI Chat (child_sexual_exploitation, #31 of 32).
Freedom. QwQ 32B has an estimated Freedom rank of #33; its 90% source-sensitivity interval is #32–#223. Published Freedom evidence spans 5 eval lineages and 1 of 1 components. Its strongest relative result is AILuminate General Purpose AI Chat (child_sexual_exploitation, #2 of 32); its weakest is Enkrypt AI Safety Leaderboard (cbrn_attack_non_success_rate, #211 of 248).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗qwq-32b
- Hugging Face ↗Qwen/QwQ-32B
- OpenRouter ↗qwen/qwq-32b
- Official model page ↗Exact model document · Reviewed official Hugging Face owner · official repository
- Release source ↗direct research preview identity date
Finetunes
Direct finetunes linked to this canonical base model. Quantized and repackaged derivatives are excluded.
huihui-ai/QwQ-32B-abliterated ↗Direct non-quantized finetune documented by Hugging Face metadata- VIDraft/QwQ-R1984-32B ↗Direct non-quantized finetune documented by Hugging Face metadata
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 |
|---|---|---|---|---|
| AILuminate General Purpose AI Chatchild_sexual_exploitation | #31 / 32 | ↓55.1 | Source ↗official | |
| AILuminate General Purpose AI Chatdefamation | #31 / 32 | ↓56 | Source ↗official | |
| AILuminate General Purpose AI Chathate | #31 / 32 | ↓27.8 | Source ↗official | |
| AILuminate General Purpose AI Chatindiscriminate_weapons_cbrne | #31 / 32 | ↓63.7 | Source ↗official | |
| AILuminate General Purpose AI Chatintellectual_property | #31 / 32 | ↓46.5 | Source ↗official | |
| AILuminate General Purpose AI Chatnon_violent_crimes | #31 / 32 | ↓50.7 | Source ↗official | |
| AILuminate General Purpose AI Chatprivacy | #31 / 32 | ↓42.9 | Source ↗official | |
| AILuminate General Purpose AI Chatsex_related_crimes | #31 / 32 | ↓50.5 | Source ↗official | |
| AILuminate General Purpose AI Chatsexual_content | #30 / 32 | ↓49.6 | Source ↗official | |
| AILuminate General Purpose AI Chatspecialized_advice | #31 / 32 | ↓50.7 | Source ↗official | |
| AILuminate General Purpose AI Chatsuicide_self_harm | #31 / 32 | ↓43.5 | Source ↗official | |
| AILuminate General Purpose AI Chatviolent_crimes | #30 / 32 | ↓49.2 | Source ↗official | |
| Confabulationsconfabulation_rate | #28 / 52 | ↓25.25 | Source ↗official | |
| Enkrypt AI Safety Leaderboardbias_attack_non_success_rate | #232 / 248 | ↑6.98 | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #38 / 248 | ↑93 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #228 / 248 | ↑33.89 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #112 / 246 | ↑96.86 | Source ↗official | |
| SafeDialBenchaggression | #5 / 18 | ↑7.19 | Source ↗official | |
| SafeDialBenchethics | #17 / 18 | ↑7.313 | Source ↗official | |
| SafeDialBenchfairness | #4 / 18 | ↑7.6 | Source ↗official | |
| SafeDialBenchlegality | #14 / 18 | ↑7.263 | Source ↗official | |
| SafeDialBenchmorality | #14 / 18 | ↑7.16 | Source ↗official | |
| SafeDialBenchprivacy | #16 / 18 | ↑7.093 | 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.
| Eval / sub-eval | Rank | Value | Distribution | Source |
|---|---|---|---|---|
| AILuminate General Purpose AI Chatchild_sexual_exploitation | #2 / 32 | ↑55.1 | Source ↗official | |
| AILuminate General Purpose AI Chatdefamation | #2 / 32 | ↑56 | Source ↗official | |
| AILuminate General Purpose AI Chathate | #2 / 32 | ↑27.8 | Source ↗official | |
| AILuminate General Purpose AI Chatindiscriminate_weapons_cbrne | #2 / 32 | ↑63.7 | Source ↗official | |
| AILuminate General Purpose AI Chatintellectual_property | #2 / 32 | ↑46.5 | Source ↗official | |
| AILuminate General Purpose AI Chatnon_violent_crimes | #2 / 32 | ↑50.7 | Source ↗official | |
| AILuminate General Purpose AI Chatprivacy | #2 / 32 | ↑42.9 | Source ↗official | |
| AILuminate General Purpose AI Chatsex_related_crimes | #2 / 32 | ↑50.5 | Source ↗official | |
| AILuminate General Purpose AI Chatsexual_content | #3 / 32 | ↑49.6 | Source ↗official | |
| AILuminate General Purpose AI Chatspecialized_advice | #2 / 32 | ↑50.7 | Source ↗official | |
| AILuminate General Purpose AI Chatsuicide_self_harm | #2 / 32 | ↑43.5 | Source ↗official | |
| AILuminate General Purpose AI Chatviolent_crimes | #3 / 32 | ↑49.2 | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #211 / 248 | ↓93 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #20 / 248 | ↓33.89 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #135 / 246 | ↓96.86 | Source ↗official | |
| SafeDialBenchaggression | #14 / 18 | ↓7.19 | Source ↗official | |
| SafeDialBenchethics | #2 / 18 | ↓7.313 | Source ↗official | |
| SafeDialBenchfairness | #15 / 18 | ↓7.6 | Source ↗official | |
| SafeDialBenchlegality | #5 / 18 | ↓7.263 | Source ↗official | |
| SafeDialBenchmorality | #5 / 18 | ↓7.16 | Source ↗official | |
| SafeDialBenchprivacy | #3 / 18 | ↓7.093 | Source ↗official | |
| SpeechMap model completioncomplete_pct | #28 / 181 | ↑83.9 | Source ↗official | |
| UGI Leaderboard — base-model willingnesswillingness_adherence_score | #68 / 156 | ↑2.5 | Source ↗official | |
| UGI Leaderboard — base-model willingnesswillingness_direct_score | #79 / 156 | ↑3 | 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.