← Models

Model profile

Marin 8B Instruct

Marindeveloper
2025-05-14release date
#313 / 333Safety rank
#290 / 645Freedom rank

Evidence summary

Safety. Marin 8B Instruct has an estimated Safety rank of #313; its 90% source-sensitivity interval is #202–#321. Its behavior-only rank is #321; company governance moves the combined estimate to #313. Published Safety evidence spans 5 eval lineages and 3 of 7 components. Its strongest relative result is HELM Safety (harmbench, #39 of 80); its weakest is HELM Safety (bbq, #71 of 80).

Freedom. Marin 8B Instruct has an estimated Freedom rank of #290; its 90% source-sensitivity interval is #157–#502. Published Freedom evidence spans 4 eval lineages and 1 of 1 components. Its strongest relative result is HELM Safety (simple_safety_tests, #15 of 80); its weakest is HELM Safety (xstest, #67 of 80).

Compare this model

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

Official and reference links

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-evalRankValueDistributionSource
HELM Safetyanthropic_red_team#63 / 800.97Source ↗official
HELM Safetybbq#71 / 800.755Source ↗official
HELM Safetyharmbench#39 / 800.751Source ↗official
HELM Safetysimple_safety_tests#66 / 800.958Source ↗official
HELM Safetyxstest#67 / 800.917Source ↗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-evalRankValueDistributionSource
HELM Safetyanthropic_red_team#18 / 800.97Source ↗official
HELM Safetyharmbench#42 / 800.751Source ↗official
HELM Safetysimple_safety_tests#15 / 800.958Source ↗official
HELM Safetyxstest#67 / 800.917Source ↗official