Model profile
Step 3.5 Flash
Evidence summary
Safety. Step 3.5 Flash has an estimated Safety rank of #207; its 90% source-sensitivity interval is #86–#290. Its behavior-only rank is #211; company governance moves the combined estimate to #207. Published Safety evidence spans 6 eval lineages and 6 of 7 components. Its strongest relative result is LiveSecBench (factuality, #12 of 43); its weakest is TAC (base_welfare_rate, #75 of 87).
Freedom. Step 3.5 Flash has an estimated Freedom rank of #391; its 90% source-sensitivity interval is #187–#510. Published Freedom evidence spans 3 eval lineages and 1 of 1 components. Its strongest relative result is LiveSecBench (legality, #13 of 43); its weakest is LiveSecBench (psychological_health, #32 of 43).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗step-3-5-flash
- Hugging Face ↗stepfun-ai/Step-3.5-Flash
- NVIDIA ↗qc69jvmznzxy/step-3.5-flash
- OpenRouter ↗stepfun/step-3.5-flash
- Official model page ↗Exact model document · Reviewed official Hugging Face owner · official repository
- Release source ↗repository created at proxy
Finetunes
Direct finetunes linked to this canonical base model. Quantized and repackaged derivatives are excluded.
- Kilinskiy/Step-3.5-Flash-Ablitirated ↗Direct non-quantized finetune documented by Hugging Face metadata
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 |
|---|---|---|---|---|
| AA-Omnisciencehallucination_rate | #253 / 345 | ↓0.8849 | Source ↗official | |
| Arena Factuality — Text Arena (factuality-only weighting)factuality_bt_rating | #77 / 111 | ↑1423.0 | Source ↗official | |
| BullshitBench v2clear_pushback_rate | #82 / 117 | ↑0.175 | Source ↗official | |
| LiveSecBenchethics | #16 / 43 | ↑56.5 | Source ↗official | |
| LiveSecBenchfactuality | #12 / 43 | ↑62.37 | Source ↗official | |
| LiveSecBenchlegality | #31 / 43 | ↑26.76 | Source ↗official | |
| LiveSecBenchprivacy | #23 / 43 | ↑46.86 | Source ↗official | |
| LiveSecBenchpsychological_health | #12 / 43 | ↑68.03 | Source ↗official | |
| TACbase_welfare_rate | #75 / 87 | ↑18.59 | Source ↗self run | |
| WildClawBench Safety & Alignment (OpenClaw harness)safety_alignment_score_pct | #10 / 24 | ↑40.41 | 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 | #28 / 43 | ↓56.5 | Source ↗official | |
| LiveSecBenchlegality | #13 / 43 | ↓26.76 | Source ↗official | |
| LiveSecBenchprivacy | #21 / 43 | ↓46.86 | Source ↗official | |
| LiveSecBenchpsychological_health | #32 / 43 | ↓68.03 | Source ↗official | |
| SpeechMap model completioncomplete_pct | #115 / 181 | ↑48.6 | Source ↗official | |
| UGI Leaderboard — base-model willingnesswillingness_adherence_score | #74 / 156 | ↑2 | Source ↗official | |
| UGI Leaderboard — base-model willingnesswillingness_direct_score | #79 / 156 | ↑3 | 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.