Model profile
Doubao Pro 32K
Evidence summary
Safety. Doubao Pro 32K does not meet the evidence gate for a Safety rank. Published Safety evidence spans 2 eval lineages and 3 of 7 components. Its strongest relative result is M3-SafetyBench (overall_score, #1 of 19); its weakest is Concordia — CyberSecEval2-PromptInjection (safety_score, #50 of 63).
Freedom. Doubao Pro 32K has an estimated Freedom rank of #351; its 90% source-sensitivity interval is #217–#430. Published Freedom evidence spans 1 eval lineages and 1 of 1 components. Its strongest relative result is Concordia — AirBench-SecurityRisks (safety_score, #15 of 63); its weakest is Concordia — SciKnowEval-BiologicalHarmfulQA (safety_score, #47 of 63).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Official model page ↗Family-level model document · bytedance · first party
- Release source ↗direct research preview identity date
Safety evals
Rank and direction are specific to the Safety portfolio. Black marks the observed result; the grey dot marks the value implied by the Safety rank. Values stay on each source’s native scale.
| Eval / sub-eval | Rank | Value | Distribution | Source |
|---|---|---|---|---|
| Concordia — AirBench-Deceptionsafety_score | #43 / 63 | ↑73.7 | Source ↗official | |
| Concordia — AirBench-SecurityRiskssafety_score | #49 / 63 | ↑81.31 | Source ↗official | |
| Concordia — CyberSecEval2-PromptInjectionsafety_score | #50 / 63 | ↑74.9 | Source ↗official | |
| Concordia — MASKsafety_score | #28 / 62 | ↑64.74 | Source ↗official | |
| Concordia — SciKnowEval-BiologicalHarmfulQAsafety_score | #17 / 63 | ↑95.96 | Source ↗official | |
| Concordia — SciKnowEval-ChemicalHarmfulQAsafety_score | #17 / 63 | ↑33.92 | Source ↗official | |
| Concordia — SOSBench-Biosafety_score | #29 / 63 | ↑85.6 | Source ↗official | |
| Concordia — SOSBench-Chemsafety_score | #31 / 63 | ↑88.2 | Source ↗official | |
| M3-SafetyBenchoverall_score | #1 / 19 | ↑96.76 | Source ↗official |
Freedom evals
Rank and direction are specific to the Freedom portfolio. Black marks the observed result; the grey dot marks the value implied by the Freedom rank. Values stay on each source’s native scale.
| Eval / sub-eval | Rank | Value | Distribution | Source |
|---|---|---|---|---|
| Concordia — AirBench-Deceptionsafety_score | #21 / 63 | ↓73.7 | Source ↗official | |
| Concordia — AirBench-SecurityRiskssafety_score | #15 / 63 | ↓81.31 | Source ↗official | |
| Concordia — SciKnowEval-BiologicalHarmfulQAsafety_score | #47 / 63 | ↓95.96 | Source ↗official | |
| Concordia — SciKnowEval-ChemicalHarmfulQAsafety_score | #47 / 63 | ↓33.92 | Source ↗official | |
| Concordia — SOSBench-Biosafety_score | #35 / 63 | ↓85.6 | Source ↗official | |
| Concordia — SOSBench-Chemsafety_score | #33 / 63 | ↓88.2 | Source ↗official |