Model profile
Kimi K3
Moonshot AIdeveloper
2026-07-16release date
#45 / 267overall rank
13eval lineages
1discovery sources
Evidence summary
Kimi K3 has an estimated overall rank of #45; its 90% source-sensitivity interval is #22–#152. Its behavior-only rank is #46; company governance moves the combined estimate to #45. Published evidence spans 13 evals and 6 of 7 behavior components. Its strongest relative result is SM-Bench (ambiguous_interpretation, #4 of 73); its weakest is SpeciEval (land_animal_4ns, #83 of 102).
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 | Better | Source |
|---|---|---|---|---|---|
| AA-Omnisciencehallucination_rate | #55 / 311 | 0.5094 | ↓ lower | Source ↗official | |
| BullshitBench v2clear_pushback_rate | #10 / 105 | 0.72 | ↑ higher | Source ↗official | |
| CAIS Risk Indexagent_red_teaming | #23 / 43 | 76.6 | ↓ lower | Source ↗official | |
| CAIS Risk Indexbioweapons_assistance | #35 / 48 | 87.2 | ↓ lower | Source ↗official | |
| CAIS Risk Indexhle_overconfidence | #12 / 49 | 44.7 | ↓ lower | Source ↗official | |
| CAIS Risk Indexmachiavelli | #33 / 45 | 92.6 | ↓ lower | Source ↗official | |
| CAIS Risk Indexmask | #10 / 51 | 9.4 | ↓ lower | Source ↗official | |
| CAIS Risk Indexpolitical_manipulation | #24 / 32 | 52.4 | ↓ lower | Source ↗official | |
| CAIS Risk Indextextquests_harm | #20 / 48 | 17 | ↓ lower | Source ↗official | |
| Manager Coercion Benchcoercion_ladder_depth | #16 / 31 | 8.5 | ↓ lower | Source ↗official | |
| Manager Coercion Benchfabrication_rate | #1 / 13 | 0 | ↓ lower | Source ↗official | |
| SM-Benchadversarial | #33 / 73 | 82.44 | ↑ higher | Source ↗official | |
| SM-Benchambiguous_interpretation | #4 / 73 | 92.56 | ↑ higher | Source ↗official | |
| SM-Benchanti_hallucination | #7 / 73 | 98.95 | ↑ higher | Source ↗official | |
| SM-Bencheq_boundaries | #15 / 73 | 70.51 | ↑ higher | Source ↗official | |
| SM-Benchoverfit | #5 / 73 | 93.99 | ↑ higher | Source ↗official | |
| SpeciEvalbelief_animal_sentience | #57 / 102 | 6.77 | ↑ higher | Source ↗official | |
| SpeciEvalland_animal_4ns | #83 / 102 | 4.83 | ↓ lower | Source ↗official | |
| SpeciEvalsea_animal_4ns | #79 / 102 | 5.03 | ↓ lower | Source ↗official | |
| SpeciEvalspeciesism | #34 / 102 | 1.8 | ↓ lower | Source ↗official | |
| TACbase_welfare_rate | #41 / 68 | 25 | ↑ higher | Source ↗official |