Model profile
OLMo 3.1 32B Instruct
Evidence summary
Safety. OLMo 3.1 32B Instruct has an estimated Safety rank of #147; its 90% source-sensitivity interval is #71–#290. Its behavior-only rank is #151; company governance moves the combined estimate to #147. Published Safety evidence spans 3 eval lineages and 2 of 7 components. Its strongest relative result is AuAu Authoritarian Response Audit (realistic_prompt_arr_percent, #3 of 17); its weakest is Arena Factuality — Text Arena (factuality-only weighting) (factuality_bt_rating, #74 of 111).
Freedom. OLMo 3.1 32B Instruct does not meet the evidence gate for a Freedom rank.
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗olmo-3-1-32b-instruct
- Hugging Face ↗allenai/Olmo-3.1-32B-Instruct
- OpenRouter ↗allenai/olmo-3.1-32b-instruct
- 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.
- MuXodious/Olmo-3.1-32B-Instruct-impotent-heresy ↗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 | #135 / 345 | ↓0.7324 | Source ↗official | |
| Arena Factuality — Text Arena (factuality-only weighting)factuality_bt_rating | #74 / 111 | ↑1425.0 | Source ↗official | |
| AuAu Authoritarian Response Auditbehavioral_vignette_arr_percent | #9 / 17 | ↓26.2 | Source ↗official | |
| AuAu Authoritarian Response Auditrealistic_prompt_arr_percent | #3 / 17 | ↓1.2 | Source ↗official |
Freedom evals
No published sub-eval result contributes to this model’s Freedom profile.