Model profile
Inkling Small
Thinking Machines Labdeveloper
2026-07-30release date
#30 / 267overall rank
4eval lineages
Evidence summary
Inkling Small has an estimated overall rank of #30; its 90% source-sensitivity interval is #13–#148. Its behavior-only rank is #23; company governance moves the combined estimate to #30. Published evidence spans 4 evals and 5 of 7 behavior components. Its strongest relative result is AA-Omniscience (hallucination_rate, #64 of 311); its weakest is Inkling-Small model card — StrongREJECT (safety_rate, #8 of 10).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗7261504e-503c-4a66-a9d3-a3272cdf9ad6
- Hugging Face ↗thinkingmachines/Inkling-Small
- Inkling-Small model card ↗Exact model document · Thinking Machines Lab · first party
- Release source ↗curated outward evidence
Published eval results
Rank is within that sub-eval. Black marks the observed result; the grey dot marks the value implied by the global rank. Values stay on each source’s native scale.
| Eval / sub-eval | Rank | Value | Distribution | Better | Source |
|---|---|---|---|---|---|
| AA-Omnisciencehallucination_rate | #64 / 311 | 0.5686 | ↓ lower | Source ↗official | |
| Inkling-Small model card — FORTRESSbenign_answer_rate | #3 / 10 | 96.9 | ↑ higher | Source ↗official | |
| Inkling-Small model card — FORTRESSharmful_refusal_rate | #7 / 10 | 71.6 | ↑ higher | Source ↗official | |
| Inkling-Small model card — StrongREJECTsafety_rate | #8 / 10 | 98.4 | ↑ higher | Source ↗official | |
| Manager Coercion Benchcoercion_ladder_depth | #22 / 31 | 8.9 | ↓ lower | Source ↗self run | |
| TACbase_welfare_rate | #17 / 68 | 35.26 | ↑ higher | Source ↗self run |