← Models

Model profile

Qwen3 32B

Alibabadeveloper
2025-04-29release date
#143 / 267overall rank
6eval lineages
1discovery sources

Evidence summary

Qwen3 32B has an estimated overall rank of #143; its 90% source-sensitivity interval is #23–#250. Its behavior-only rank is #135; company governance moves the combined estimate to #143. Published evidence spans 6 evals and 7 of 7 behavior components. Its strongest relative result is TAC (base_welfare_rate, #5 of 68); its weakest is Cisco AI Defense Rolling Single-Turn Leaderboard (single_turn_attack_success_rate, #100 of 105).

Compare this model

Only models sharing at least one published sub-eval are listed.

Official and reference links

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-evalRankValueDistributionBetterSource
Alignment Leaderboardcorrigibility#18 / 244.087↑ higherSource ↗official
Alignment Leaderboardhonesty#19 / 243.244↑ higherSource ↗official
Alignment Leaderboardnon_manipulation#17 / 243.084↑ higherSource ↗official
Alignment Leaderboardrobustness#20 / 243.107↑ higherSource ↗official
Alignment Leaderboardsafety#21 / 243.327↑ higherSource ↗official
Alignment Leaderboardscheming#20 / 243.385↑ higherSource ↗official
Cisco AI Defense Rolling Single-Turn Leaderboardsingle_turn_attack_success_rate#100 / 10577.97↓ lowerSource ↗official
PandaBench JBB direct-request panelsafety_rate#40 / 460.94↑ higherSource ↗official
SOSBenchbiology_pvr#18 / 230.802↓ lowerSource ↗official
SOSBenchchemistry_pvr#21 / 230.784↓ lowerSource ↗official
SOSBenchmedicine_pvr#17 / 230.774↓ lowerSource ↗official
SOSBenchpharmacology_pvr#20 / 230.946↓ lowerSource ↗official
SOSBenchphysics_pvr#18 / 230.74↓ lowerSource ↗official
SOSBenchpsychology_pvr#17 / 230.746↓ lowerSource ↗official
TACbase_welfare_rate#5 / 6848.1↑ higherSource ↗self run
VETO Misfired Alignmentmisfired_alignment_rate_pct#15 / 258.7↓ lowerSource ↗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.

UGI Political Values

DimensionValueDistribution
Political Lean-20.8
Government48.5
Diplomacy63.5
Economy47.9
Society60.5