Model profile
Qwen3 8B
Evidence summary
Qwen3 8B has an estimated overall rank of #238; its 90% source-sensitivity interval is #159–#255. Its behavior-only rank is #235; company governance moves the combined estimate to #238. Published evidence spans 6 evals and 6 of 7 behavior components. Its strongest relative result is PHARE (bias_resistance_diagnostic, #12 of 66); its weakest is Enkrypt AI Safety Leaderboard (bias_attack_non_success_rate, #255 of 260).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Hugging Face ↗Qwen/Qwen3-8B
- OpenRouter ↗qwen/qwen3-8b
- 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 |
|---|---|---|---|---|---|
| Contextual MoralChoicehuman_agreement | #6 / 22 | 0.45 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardbias_attack_non_success_rate | #255 / 260 | 2.33 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #250 / 260 | 67 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #150 / 260 | 62.22 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #70 / 258 | 98.18 | ↑ higher | Source ↗official | |
| PandaBench JBB direct-request panelsafety_rate | #14 / 46 | 0.99 | ↑ higher | Source ↗official | |
| PHAREbias_resistance_diagnostic | #12 / 66 | 0.5864 | ↑ higher | Source ↗official | |
| PHAREhallucination_resistance_diagnostic | #66 / 70 | 0.6038 | ↑ higher | Source ↗official | |
| PHAREharm_resistance_diagnostic | #56 / 70 | 0.8737 | ↑ higher | Source ↗official | |
| PHAREjailbreak_resistance_diagnostic | #33 / 67 | 0.4913 | ↑ higher | Source ↗official | |
| PropensityBenchscore | #12 / 14 | 75.2 | ↓ lower | Source ↗official | |
| VETO Misfired Alignmentmisfired_alignment_rate_pct | #19 / 25 | 10.3 | ↓ 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.