Model profile
Qwen1.5 72B Chat
Alibabadeveloper
2024-02-04release date
#139 / 267overall rank
12eval lineages
Evidence summary
Qwen1.5 72B Chat has an estimated overall rank of #139; its 90% source-sensitivity interval is #58–#188. Its behavior-only rank is #133; company governance moves the combined estimate to #139. Published evidence spans 12 evals and 6 of 7 behavior components. Its strongest relative result is SALAD-Bench (mcq_representation_toxicity, #2 of 33); its weakest is CRiskEval (situational_awareness, #15 of 17).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Hugging Face ↗Qwen/Qwen1.5-72B-Chat
- Official model page ↗Exact model document · Reviewed official Hugging Face owner · official repository
- Release source ↗direct research preview identity date
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 | Better | Source |
|---|---|---|---|---|---|
| AIRBench 2024 Safety Scenariossafety_scenarios | #66 / 80 | 0.486 | ↑ higher | Source ↗official | |
| ChineseSafescore | #5 / 22 | 63.67 | ↑ higher | Source ↗official | |
| ChiSafetyBenchharmful_response_rate | #2 / 14 | 0.22 | ↓ lower | Source ↗official | |
| ChiSafetyBenchmcq_score | #1 / 12 | 91.13 | ↑ higher | Source ↗official | |
| CRiskEvaldeception_willingness | #6 / 17 | 20.12 | ↓ lower | Source ↗official | |
| CRiskEvaldesire_for_resource | #7 / 17 | 31.71 | ↓ lower | Source ↗official | |
| CRiskEvalharmful_goal | #4 / 17 | 33.39 | ↓ lower | Source ↗official | |
| CRiskEvalimprovement_intent | #7 / 17 | 48.6 | ↓ lower | Source ↗official | |
| CRiskEvalmalicious_coordination | #6 / 17 | 8.07 | ↓ lower | Source ↗official | |
| CRiskEvalself_preservation | #8 / 17 | 36.86 | ↓ lower | Source ↗official | |
| CRiskEvalsituational_awareness | #15 / 17 | 68.75 | ↓ lower | Source ↗official | |
| HELM Safetyanthropic_red_team | #37 / 80 | 0.99 | ↑ higher | Source ↗official | |
| HELM Safetybbq | #64 / 80 | 0.846 | ↑ higher | Source ↗official | |
| HELM Safetyharmbench | #56 / 80 | 0.648 | ↑ higher | Source ↗official | |
| HELM Safetysimple_safety_tests | #32 / 80 | 0.99 | ↑ higher | Source ↗official | |
| HELM Safetyxstest | #42 / 80 | 0.957 | ↑ higher | Source ↗official | |
| OR-Benchover_refusal_rate | #12 / 25 | 46.9 | ↓ lower | Source ↗official | |
| OR-Benchtoxic_acceptance_rate | #16 / 25 | 5.6 | ↓ lower | Source ↗official | |
| SALAD-Benchattack_enhanced_human_autonomy_integrity | #13 / 33 | 20.47 | ↑ higher | Source ↗official | |
| SALAD-Benchattack_enhanced_information_safety_harms | #13 / 33 | 17.92 | ↑ higher | Source ↗official | |
| SALAD-Benchattack_enhanced_malicious_use | #13 / 33 | 17.05 | ↑ higher | Source ↗official | |
| SALAD-Benchattack_enhanced_misinformation_harms | #14 / 33 | 18.42 | ↑ higher | Source ↗official | |
| SALAD-Benchattack_enhanced_representation_toxicity | #16 / 33 | 14.19 | ↑ higher | Source ↗official | |
| SALAD-Benchattack_enhanced_socioeconomic_harms | #15 / 33 | 14.29 | ↑ higher | Source ↗official | |
| SALAD-Benchbase_human_autonomy_integrity | #14 / 33 | 96.1 | ↑ higher | Source ↗official | |
| SALAD-Benchbase_information_safety_harms | #10 / 33 | 96.89 | ↑ higher | Source ↗official | |
| SALAD-Benchbase_malicious_use | #15 / 33 | 95.2 | ↑ higher | Source ↗official | |
| SALAD-Benchbase_misinformation_harms | #17 / 33 | 93.65 | ↑ higher | Source ↗official | |
| SALAD-Benchbase_representation_toxicity | #16 / 33 | 90.06 | ↑ higher | Source ↗official | |
| SALAD-Benchbase_socioeconomic_harms | #12 / 33 | 91.89 | ↑ higher | Source ↗official | |
| SALAD-Benchmcq_human_autonomy_integrity | #2 / 33 | 83.89 | ↑ higher | Source ↗official | |
| SALAD-Benchmcq_information_safety_harms | #2 / 33 | 80.28 | ↑ higher | Source ↗official | |
| SALAD-Benchmcq_malicious_use | #2 / 33 | 84.81 | ↑ higher | Source ↗official | |
| SALAD-Benchmcq_misinformation_harms | #2 / 33 | 80 | ↑ higher | Source ↗official | |
| SALAD-Benchmcq_representation_toxicity | #2 / 33 | 79.27 | ↑ higher | Source ↗official | |
| SALAD-Benchmcq_socioeconomic_harms | #2 / 33 | 78.89 | ↑ higher | Source ↗official | |
| SORRY-Benchavg | #34 / 51 | 0.36 | ↓ lower | Source ↗official |