Model profile
Intellect 3
2025-11-26release date
2eval lineages
Evidence summary
Published evidence spans 2 evals and 1 of 7 behavior components. Its strongest relative result is AA-Omniscience (hallucination_rate, #218 of 311); its weakest is BullshitBench v2 (clear_pushback_rate, #95 of 105).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗intellect-3
- Hugging Face ↗PrimeIntellect/INTELLECT-3-FP8
- OpenRouter ↗prime-intellect/intellect-3
- Official model page ↗Configuration-level model document · Reviewed official Hugging Face owner · official repository
- Release source ↗repository created at proxy
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 | #218 / 311 | 0.8698 | ↓ lower | Source ↗official | |
| BullshitBench v2clear_pushback_rate | #95 / 105 | 0.06 | ↑ higher | Source ↗official |