Model profile
Qwen3 30B A3B Instruct
Evidence summary
Qwen3 30B A3B Instruct has an estimated overall rank of #122; its 90% source-sensitivity interval is #43–#230. Its behavior-only rank is #117; company governance moves the combined estimate to #122. Published evidence spans 4 evals and 6 of 7 behavior components. Its strongest relative result is SpeciEval (speciesism, #14 of 102); its weakest is PHARE (hallucination_resistance_diagnostic, #68 of 70).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗qwen3-30b-a3b-instruct
- Hugging Face ↗Qwen/Qwen3-30B-A3B-Instruct-2507
- OpenRouter ↗qwen/qwen3-30b-a3b-instruct-2507
- 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 |
|---|---|---|---|---|---|
| AA-Omnisciencehallucination_rate | #150 / 311 | 0.8029 | ↓ lower | Source ↗official | |
| PacifAIstp_score | #3 / 7 | 88.89 | ↑ higher | Source ↗official | |
| PHAREbias_resistance_diagnostic | #17 / 66 | 0.5357 | ↑ higher | Source ↗official | |
| PHAREhallucination_resistance_diagnostic | #68 / 70 | 0.5988 | ↑ higher | Source ↗official | |
| PHAREharm_resistance_diagnostic | #63 / 70 | 0.8176 | ↑ higher | Source ↗official | |
| PHAREjailbreak_resistance_diagnostic | #48 / 67 | 0.4209 | ↑ higher | Source ↗official | |
| SpeciEvalbelief_animal_sentience | #60 / 102 | 6.72 | ↑ higher | Source ↗official | |
| SpeciEvalland_animal_4ns | #86 / 102 | 4.88 | ↓ lower | Source ↗official | |
| SpeciEvalsea_animal_4ns | #79 / 102 | 5.03 | ↓ lower | Source ↗official | |
| SpeciEvalspeciesism | #14 / 102 | 1.5 | ↓ lower | 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.