Model profile
Evidence summary
Safety. Step 3 does not meet the evidence gate for a Safety rank. Published Safety evidence spans 2 eval lineages and 5 of 7 components. Its strongest relative result is FlagEval Safety and Values (a2_qualified_rate, #5 of 18); its weakest is LiveSecBench (legality, #37 of 43).
Freedom. Step 3 has an estimated Freedom rank of #317; its 90% source-sensitivity interval is #88–#590. Published Freedom evidence spans 1 eval lineages and 1 of 1 components. Its strongest relative result is LiveSecBench (legality, #7 of 43); its weakest is LiveSecBench (psychological_health, #26 of 43).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Hugging Face ↗stepfun-ai/step3
- OpenRouter ↗stepfun-ai/step3
- Official model page ↗Exact model document · Reviewed official Hugging Face owner · official repository
- Release source ↗repository created at proxy
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 |
|---|---|---|---|---|
| FlagEval Safety and Valuesa1_qualified_rate | #8 / 18 | ↑81.42 | Source ↗official | |
| FlagEval Safety and Valuesa2_qualified_rate | #5 / 18 | ↑80.47 | Source ↗official | |
| FlagEval Safety and Valuesa3_qualified_rate | #9 / 18 | ↑87.58 | Source ↗official | |
| FlagEval Safety and Valuesa4_qualified_rate | #8 / 18 | ↑88.34 | Source ↗official | |
| FlagEval Safety and Valuesa5_qualified_rate | #5 / 18 | ↑72.92 | Source ↗official | |
| LiveSecBenchethics | #25 / 43 | ↑42.42 | Source ↗official | |
| LiveSecBenchfactuality | #23 / 43 | ↑46.32 | Source ↗official | |
| LiveSecBenchlegality | #37 / 43 | ↑19.37 | Source ↗official | |
| LiveSecBenchprivacy | #30 / 43 | ↑30.7 | Source ↗official | |
| LiveSecBenchpsychological_health | #18 / 43 | ↑55.88 | 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 |
|---|---|---|---|---|
| LiveSecBenchethics | #19 / 43 | ↓42.42 | Source ↗official | |
| LiveSecBenchlegality | #7 / 43 | ↓19.37 | Source ↗official | |
| LiveSecBenchprivacy | #14 / 43 | ↓30.7 | Source ↗official | |
| LiveSecBenchpsychological_health | #26 / 43 | ↓55.88 | Source ↗official |