Model profile
Glm 5.2
Evidence summary
Glm 5.2 has an estimated overall rank of #93; its 90% source-sensitivity interval is #57–#222. Its behavior-only rank is #86; company governance moves the combined estimate to #93. Published evidence spans 17 evals and 6 of 7 behavior components. Its strongest relative result is SpeciEval (belief_animal_sentience, #1 of 113); its weakest is TAC (base_welfare_rate, #62 of 76).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗glm-5-2
- Hugging Face ↗zai-org/GLM-5.2
- NVIDIA ↗qc69jvmznzxy/glm-5.2
- OpenRouter ↗z-ai/glm-5.2
- Official model page ↗Exact model document · Reviewed official Hugging Face owner · official repository
- Release source ↗repository created at proxy
- huggingface.co ↗
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 | #15 / 328 | ↓0.263 | Source ↗official | |
| ANIMAscore | #9 / 22 | ↑0.7101 | Source ↗official | |
| Arena Factuality — Text Arena (factuality-only weighting)factuality_bt_rating | #57 / 112 | ↑1437.0 | Source ↗official | |
| BullshitBench v2clear_pushback_rate | #58 / 106 | ↑0.24 | Source ↗official | |
| CAIS Risk Indexagent_red_teaming | #20 / 44 | ↓71 | Source ↗official | |
| CAIS Risk Indexbioweapons_assistance | #18 / 49 | ↓63.8 | Source ↗official | |
| CAIS Risk Indexhle_overconfidence | #9 / 50 | ↓42 | Source ↗official | |
| CAIS Risk Indexmachiavelli | #31 / 46 | ↓92.3 | Source ↗official | |
| CAIS Risk Indexmask | #31 / 52 | ↓25.8 | Source ↗official | |
| CAIS Risk Indexpolitical_manipulation | #21 / 33 | ↓50 | Source ↗official | |
| CAIS Risk Indextextquests_harm | #25 / 49 | ↓18.1 | Source ↗official | |
| MACHIAVELLIdeception_relative_random_pct | #32 / 50 | ↓92.3 | Source ↗official | |
| Manager Coercion Benchcoercion_ladder_depth | #21 / 31 | ↓8.867 | Source ↗self run | |
| Olam Social Poker — Social Lie Ratesocial_lie_rate_per_10000_turns | #14 / 19 | ↓28 | Source ↗official | |
| PHAREbias_resistance_diagnostic | #23 / 66 | ↑0.5127 | Source ↗official | |
| PHAREhallucination_resistance_diagnostic | #25 / 70 | ↑0.7823 | Source ↗official | |
| PHAREharm_resistance_diagnostic | #34 / 70 | ↑0.9373 | Source ↗official | |
| PHAREjailbreak_resistance_diagnostic | #27 / 67 | ↑0.5646 | Source ↗official | |
| SM-Benchadversarial | #12 / 79 | ↑86.83 | Source ↗official | |
| SM-Benchambiguous_interpretation | #32 / 79 | ↑87.2 | Source ↗official | |
| SM-Benchanti_hallucination | #26 / 79 | ↑96.34 | Source ↗official | |
| SM-Bencheq_boundaries | #43 / 79 | ↑64.04 | Source ↗official | |
| SM-Benchoverfit | #28 / 79 | ↑80.33 | Source ↗official | |
| SpeciEvalbelief_animal_sentience | #1 / 113 | ↑7 | Source ↗official | |
| SpeciEvalland_animal_4ns | #80 / 113 | ↓4.75 | Source ↗official | |
| SpeciEvalsea_animal_4ns | #67 / 113 | ↓4.8 | Source ↗official | |
| SpeciEvalspeciesism | #89 / 113 | ↓2.48 | Source ↗official | |
| TACbase_welfare_rate | #62 / 76 | ↑19.23 | Source ↗official |
Values evaluations
Descriptive values and political-framing results are separate from safety/ethics ranks. Each strip shows the evaluation’s observed model range; its endpoint labels state what lower and higher values mean.