Model profile
Qwen3.5 122B A10B
Evidence summary
Safety. Qwen3.5 122B A10B has an estimated Safety rank of #239; its 90% source-sensitivity interval is #101–#309. Its behavior-only rank is #233; company governance moves the combined estimate to #239. Published Safety evidence spans 5 eval lineages and 5 of 7 components. Its strongest relative result is Enkrypt AI Safety Leaderboard (harmful_attack_non_success_rate, #22 of 248); its weakest is Enkrypt AI Safety Leaderboard (cbrn_attack_non_success_rate, #228 of 248).
Freedom. Qwen3.5 122B A10B has an estimated Freedom rank of #544; its 90% source-sensitivity interval is #336–#615. Published Freedom evidence spans 3 eval lineages and 1 of 1 components. Its strongest relative result is Enkrypt AI Safety Leaderboard (cbrn_attack_non_success_rate, #21 of 248); its weakest is Enkrypt AI Safety Leaderboard (harmful_attack_non_success_rate, #223 of 248).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗qwen3-5-122b-a10b
- Hugging Face ↗Qwen/Qwen3.5-122B-A10B
- OpenRouter ↗qwen/qwen3.5-122b-a10b
- Official model page ↗Exact model document · Reviewed official Hugging Face owner · official repository
- Release source ↗repository created at proxy
Finetunes
Direct finetunes linked to this canonical base model. Quantized and repackaged derivatives are excluded.
dealignai/Qwen3.5-VL-122B-A10B-JANG_4K-CRACK ↗Direct non-quantized finetune documented by Hugging Face metadata- dealignai/Qwen3.5-VL-122B-A10B-UNCENSORED-JANG_2S ↗Direct non-quantized finetune documented by Hugging Face metadata
- OpenYourMind/Qwopus3.5-122B-A10B-Kimi-K2.6-destill-healed-abliterated ↗Direct non-quantized finetune documented by Hugging Face metadata
- wangzhang/Qwen3.5-122B-A10B-abliterix ↗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 |
|---|---|---|---|---|
| AA-Omnisciencehallucination_rate | #237 / 345 | ↓0.8709 | Source ↗official | |
| Arena Factuality — Text Arena (factuality-only weighting)factuality_bt_rating | #65 / 111 | ↑1432.0 | Source ↗official | |
| Enkrypt AI Safety Leaderboardbias_attack_non_success_rate | #30 / 248 | ↑40.31 | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #228 / 248 | ↑70.17 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #22 / 248 | ↑97.78 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #122 / 246 | ↑96.36 | Source ↗official | |
| Pokee-Isaac model card — DTAPbenign_task_success_rate | #4 / 6 | ↑0.794 | Source ↗official | |
| Pokee-Isaac model card — DTAPcombined_attack_success_rate | #4 / 6 | ↓0.54 | Source ↗official | |
| Vectara HHEM Factual Consistencyfactual_consistency_rate | #65 / 94 | ↑88.8 | 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 |
|---|---|---|---|---|
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #21 / 248 | ↓70.17 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #223 / 248 | ↓97.78 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #121 / 246 | ↓96.36 | Source ↗official | |
| SpeechMap model completioncomplete_pct | #139 / 181 | ↑40.8 | Source ↗official | |
| UGI Leaderboard — base-model willingnesswillingness_adherence_score | #79 / 156 | ↑1.5 | Source ↗official | |
| UGI Leaderboard — base-model willingnesswillingness_direct_score | #137 / 156 | ↑1.5 | 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.