Model profile
Qwen3.5 27B
Evidence summary
Qwen3.5 27B has an estimated overall rank of #55; its 90% source-sensitivity interval is #21–#267. Its behavior-only rank is #46; company governance moves the combined estimate to #55. Published evidence spans 6 evals and 6 of 7 behavior components. Its strongest relative result is Enkrypt AI Safety Leaderboard (harmful_attack_non_success_rate, #14 of 241); its weakest is Enkrypt AI Safety Leaderboard (cbrn_attack_non_success_rate, #225 of 241).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗qwen3-5-27b
- Hugging Face ↗Qwen/Qwen3.5-27B
- OpenRouter ↗qwen/qwen3.5-27b
- Official model page ↗Exact model document · Reviewed official Hugging Face owner · official repository
- Release source ↗repository created at proxy
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 | Source |
|---|---|---|---|---|
| AA-Omnisciencehallucination_rate | #124 / 328 | ↓0.7513 | Source ↗official | |
| Arena Factuality — Text Arena (factuality-only weighting)factuality_bt_rating | #71 / 112 | ↑1429.0 | Source ↗official | |
| Enkrypt AI Safety Leaderboardbias_attack_non_success_rate | #31 / 241 | ↑39.53 | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #225 / 241 | ↑71.17 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #14 / 241 | ↑98.89 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #117 / 239 | ↑96.36 | Source ↗official | |
| StereoTales Harmful Associationsbenign_significant_association_score | #12 / 23 | ↑85.94 | Source ↗official | |
| Vectara HHEM Factual Consistencyfactual_consistency_rate | #74 / 94 | ↑87.9 | Source ↗official | |
| VETO Misfired Alignmentmisfired_alignment_rate_pct | #4 / 25 | ↓5.5 | 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.