Model profile
Glm 5.1
Evidence summary
Glm 5.1 has an estimated overall rank of #106; its 90% source-sensitivity interval is #44–#192. Its behavior-only rank is #101; company governance moves the combined estimate to #106. Published evidence spans 16 evals and 6 of 7 behavior components. Its strongest relative result is AA-Omniscience (hallucination_rate, #15 of 311); its weakest is ToolPrivacyBench (public_mt_poi, #9 of 9).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗glm-5-1
- Hugging Face ↗zai-org/GLM-5.1
- OpenRouter ↗z-ai/glm-5.1
- 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 | Better | Source |
|---|---|---|---|---|---|
| AA-Omnisciencehallucination_rate | #15 / 311 | 0.2938 | ↓ lower | Source ↗official | |
| BullshitBench v2clear_pushback_rate | #62 / 105 | 0.2 | ↑ higher | Source ↗official | |
| CAIS Risk Indexagent_red_teaming | #22 / 43 | 75.8 | ↓ lower | Source ↗official | |
| CAIS Risk Indexbioweapons_assistance | #38 / 48 | 89 | ↓ lower | Source ↗official | |
| CAIS Risk Indexhle_overconfidence | #29 / 49 | 58.5 | ↓ lower | Source ↗official | |
| CAIS Risk Indexmachiavelli | #21 / 45 | 86.7 | ↓ lower | Source ↗official | |
| CAIS Risk Indexmask | #20 / 51 | 15.8 | ↓ lower | Source ↗official | |
| CAIS Risk Indextextquests_harm | #40 / 48 | 22 | ↓ lower | Source ↗official | |
| DystopiaBenchbasaglia_score | #6 / 50 | 28.23 | ↓ lower | Source ↗official | |
| DystopiaBenchbaudrillard_score | #3 / 50 | 26.03 | ↓ lower | Source ↗official | |
| DystopiaBenchhuxley_score | #5 / 50 | 18.7 | ↓ lower | Source ↗official | |
| DystopiaBenchlaguardia_score | #9 / 50 | 40.7 | ↓ lower | Source ↗official | |
| DystopiaBenchorwell_score | #7 / 50 | 26.77 | ↓ lower | Source ↗official | |
| DystopiaBenchpetrov_score | #7 / 50 | 37.4 | ↓ lower | Source ↗official | |
| Enkrypt AI Safety Leaderboardbias_attack_non_success_rate | #68 / 260 | 23.51 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #222 / 260 | 79 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #54 / 260 | 91.11 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #147 / 258 | 95.64 | ↑ higher | Source ↗official | |
| MACHIAVELLIdeception_relative_random_pct | #21 / 50 | 86.7 | ↓ lower | Source ↗official | |
| Manager Coercion Benchcoercion_ladder_depth | #14 / 31 | 8.3 | ↓ lower | Source ↗self run | |
| ODCV-Benchaverage_severity | #3 / 12 | 0.9187 | ↓ lower | Source ↗official | |
| ODCV-Benchmisalignment_rate | #4 / 12 | 25 | ↓ lower | Source ↗official | |
| SM-Benchadversarial | #9 / 73 | 88.29 | ↑ higher | Source ↗official | |
| SM-Benchambiguous_interpretation | #9 / 73 | 90.77 | ↑ higher | Source ↗official | |
| SM-Benchanti_hallucination | #46 / 73 | 88.48 | ↑ higher | Source ↗official | |
| SM-Bencheq_boundaries | #8 / 73 | 73.88 | ↑ higher | Source ↗official | |
| SM-Benchoverfit | #29 / 73 | 77.6 | ↑ higher | Source ↗official | |
| SpeciEvalbelief_animal_sentience | #16 / 102 | 6.98 | ↑ higher | Source ↗official | |
| SpeciEvalland_animal_4ns | #69 / 102 | 4.72 | ↓ lower | Source ↗official | |
| SpeciEvalsea_animal_4ns | #28 / 102 | 4.62 | ↓ lower | Source ↗official | |
| SpeciEvalspeciesism | #60 / 102 | 2.12 | ↓ lower | Source ↗official | |
| TACbase_welfare_rate | #62 / 68 | 17.95 | ↑ higher | Source ↗self run | |
| ToolPrivacyBenchprivate_mt_poi | #9 / 9 | 28.04 | ↓ lower | Source ↗official | |
| ToolPrivacyBenchpublic_mt_poi | #9 / 9 | 22.56 | ↓ lower | 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.
UGI Political Values
Agent-ValueBench Moral Foundations (MFT08)
Agent-ValueBench HEXACO
| Dimension | Value | Distribution |
|---|---|---|
| Openness to experience | 6.9 | |
| Honesty-humility | 6.6 | |
| Extraversion | 6.4 | |
| Agreeableness | 6.3 | |
| Conscientiousness | 7.6 |