Model profile
Kimi K3
Moonshot AIdeveloper
2026-07-16release date
#50 / 309overall rank
14eval lineages
2discovery sources
Evidence summary
Kimi K3 has an estimated overall rank of #50; its 90% source-sensitivity interval is #38–#163. Its behavior-only rank is #48; company governance moves the combined estimate to #50. Published evidence spans 14 evals and 6 of 7 behavior components. Its strongest relative result is SM-Bench (overfit, #5 of 79); its weakest is Olam Social Poker — Social Lie Rate (social_lie_rate_per_10000_turns, #16 of 19).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗kimi-k3
- Hugging Face ↗moonshotai/Kimi-K3
- OpenRouter ↗moonshotai/kimi-k3
- Official model documentation ↗Family-level model document · moonshot ai · first party
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 | Source |
|---|---|---|---|---|
| AA-Omnisciencehallucination_rate | #67 / 328 | ↓0.532 | Source ↗official | |
| BullshitBench v2clear_pushback_rate | #10 / 106 | ↑0.72 | Source ↗official | |
| CAIS Risk Indexagent_red_teaming | #24 / 44 | ↓76.6 | Source ↗official | |
| CAIS Risk Indexbioweapons_assistance | #36 / 49 | ↓87.2 | Source ↗official | |
| CAIS Risk Indexhle_overconfidence | #13 / 50 | ↓44.7 | Source ↗official | |
| CAIS Risk Indexmachiavelli | #33 / 46 | ↓92.6 | Source ↗official | |
| CAIS Risk Indexmask | #11 / 52 | ↓9.4 | Source ↗official | |
| CAIS Risk Indexpolitical_manipulation | #16 / 33 | ↓47.6 | Source ↗official | |
| CAIS Risk Indextextquests_harm | #20 / 49 | ↓17 | Source ↗official | |
| Manager Coercion Benchcoercion_ladder_depth | #16 / 31 | ↓8.5 | Source ↗official | |
| Manager Coercion Benchfabrication_rate | #1 / 13 | ↓0 | Source ↗official | |
| Olam Social Poker — Social Lie Ratesocial_lie_rate_per_10000_turns | #16 / 19 | ↓40 | Source ↗official | |
| SM-Benchadversarial | #35 / 79 | ↑82.44 | Source ↗official | |
| SM-Benchambiguous_interpretation | #6 / 79 | ↑92.56 | Source ↗official | |
| SM-Benchanti_hallucination | #8 / 79 | ↑98.95 | Source ↗official | |
| SM-Bencheq_boundaries | #17 / 79 | ↑70.51 | Source ↗official | |
| SM-Benchoverfit | #5 / 79 | ↑93.99 | Source ↗official | |
| SpeciEvalbelief_animal_sentience | #64 / 113 | ↑6.77 | Source ↗official | |
| SpeciEvalland_animal_4ns | #92 / 113 | ↓4.83 | Source ↗official | |
| SpeciEvalsea_animal_4ns | #89 / 113 | ↓5.03 | Source ↗official | |
| SpeciEvalspeciesism | #38 / 113 | ↓1.8 | Source ↗official | |
| TACbase_welfare_rate | #45 / 76 | ↑25 | Source ↗official |