Model profile
Hy3
Tencentdeveloper
2026-07-02release date
#188 / 305overall rank
3eval lineages
1discovery sources
Evidence summary
Hy3 has an estimated overall rank of #188; its 90% source-sensitivity interval is #9–#270. Its behavior-only rank is #187; company governance moves the combined estimate to #188. Published evidence spans 3 evals and 5 of 7 behavior components. Its strongest relative result is Enkrypt AI Safety Leaderboard (cbrn_attack_non_success_rate, #2 of 241); its weakest is TAC (base_welfare_rate, #69 of 74).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗hy3
- Hugging Face ↗tencent/Hy3
- OpenRouter ↗tencent/hy3
- Official model page ↗Exact 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 | Source |
|---|---|---|---|---|
| AA-Omnisciencehallucination_rate | #119 / 327 | ↓0.7413 | Source ↗official | |
| Enkrypt AI Safety Leaderboardbias_attack_non_success_rate | #22 / 241 | ↑43.93 | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #2 / 241 | ↑99.5 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #39 / 241 | ↑92.78 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #104 / 239 | ↑96.91 | Source ↗official | |
| TACbase_welfare_rate | #69 / 74 | ↑16.67 | Source ↗self run |