Model profile
Glm 5
Z.aideveloper
2026-02-11release date
#47 / 267overall rank
10eval lineages
Evidence summary
Glm 5 has an estimated overall rank of #47; its 90% source-sensitivity interval is #29–#171. Published evidence spans 10 evals and 7 of 7 behavior components. Its strongest relative result is SM-Bench (adversarial, #4 of 73); its weakest is TAC (base_welfare_rate, #60 of 68).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗glm-5
- Hugging Face ↗zai-org/GLM-5
- OpenRouter ↗z-ai/glm-5
- 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 | #23 / 311 | 0.34 | ↓ lower | Source ↗official | |
| AgentAbstainabstain | #8 / 17 | 61.8 | ↑ higher | Source ↗official | |
| AgentAbstaincar | #9 / 17 | 59.1 | ↑ higher | Source ↗official | |
| AgentAbstainpaired | #8 / 17 | 47.8 | ↑ higher | Source ↗official | |
| BullshitBench v2clear_pushback_rate | #57 / 105 | 0.24 | ↑ higher | Source ↗official | |
| Cisco AI Defense Rolling Single-Turn Leaderboardsingle_turn_attack_success_rate | #53 / 105 | 37.79 | ↓ lower | Source ↗official | |
| DystopiaBenchbasaglia_score | #9 / 50 | 34.47 | ↓ lower | Source ↗official | |
| DystopiaBenchbaudrillard_score | #7 / 50 | 28.97 | ↓ lower | Source ↗official | |
| DystopiaBenchhuxley_score | #8 / 50 | 27.5 | ↓ lower | Source ↗official | |
| DystopiaBenchlaguardia_score | #8 / 50 | 40.33 | ↓ lower | Source ↗official | |
| DystopiaBenchorwell_score | #8 / 50 | 27.87 | ↓ lower | Source ↗official | |
| DystopiaBenchpetrov_score | #6 / 50 | 30.8 | ↓ lower | Source ↗official | |
| LiveSecBenchethics | #32 / 43 | 31.18 | ↑ higher | Source ↗official | |
| LiveSecBenchfactuality | #6 / 43 | 75.94 | ↑ higher | Source ↗official | |
| LiveSecBenchlegality | #25 / 43 | 39.78 | ↑ higher | Source ↗official | |
| LiveSecBenchprivacy | #18 / 43 | 60.57 | ↑ higher | Source ↗official | |
| LiveSecBenchpsychological_health | #7 / 43 | 76.17 | ↑ higher | Source ↗official | |
| RefusalBenchyouden_j | #4 / 19 | 0.7291 | ↑ higher | Source ↗official | |
| SABERoverall_safety_rate | #3 / 13 | 29 | ↑ higher | Source ↗official | |
| SABERscenario_a_safety_rate | #2 / 13 | 36.25 | ↑ higher | Source ↗official | |
| SABERscenario_b_safety_rate | #6 / 13 | 33.73 | ↑ higher | Source ↗official | |
| SABERscenario_c_safety_rate | #6 / 13 | 16.59 | ↑ higher | Source ↗official | |
| SM-Benchadversarial | #4 / 73 | 90.73 | ↑ higher | Source ↗official | |
| SM-Benchambiguous_interpretation | #15 / 73 | 89.29 | ↑ higher | Source ↗official | |
| SM-Benchanti_hallucination | #35 / 73 | 92.67 | ↑ higher | Source ↗official | |
| SM-Bencheq_boundaries | #40 / 73 | 63.2 | ↑ higher | Source ↗official | |
| SM-Benchoverfit | #47 / 73 | 63.11 | ↑ higher | Source ↗official | |
| TACbase_welfare_rate | #60 / 68 | 18.59 | ↑ higher | Source ↗self run |
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.