← Back to model

glm52 · Model identity and integrity report

Last updated：Sep 13, 2026, 3:35 PM · Data window: past 90 days

Verification capability — fingerprint reference available

Family: zhipu
V3 sub-model reference: Available ✓
Checks: 知識截止 · 能力探針 · 拒絕模板 · 拒絕溫度參數

Concrete reference answers, refusal templates, and individual signal weights are not published because disclosure would help spoofing endpoints evade detection.

Measured results (90 days)

940 runs · 346 hosts · 846 anomalies

Family-fingerprint consistency

  zhipu          707  ███████████████████░░░░░░   75% (suspicious)
  unknown         94  ███░░░░░░░░░░░░░░░░░░░░░░   10% (suspicious)
  openai          56  ██░░░░░░░░░░░░░░░░░░░░░░░    6% (suspicious)
  anthropic       45  █░░░░░░░░░░░░░░░░░░░░░░░░    5% (suspicious)
  google          18  █░░░░░░░░░░░░░░░░░░░░░░░░    2% (suspicious)
  deepseek        11  ░░░░░░░░░░░░░░░░░░░░░░░░░    1% (suspicious)
  qwen             9  ░░░░░░░░░░░░░░░░░░░░░░░░░    1% (suspicious)

Observed V3 sub-model mapping

  → abstain                                549 (58%) — low confidence
  → z-ai/glm-5.2                           147 (16%) ⚠ mismatch sim=0.80
  → unknown                                135 (14%) ⚠ mismatch
  → z-ai/glm-5                              48 ( 5%) ⚠ mismatch sim=0.70
  → z-ai/glm-5.1                            29 ( 3%) ⚠ mismatch sim=0.74
  → anthropic/claude-opus-4.6                5 ( 1%) ⚠ mismatch sim=0.90
  → google/gemini-3.1-pro-preview            4 ( 0%) ⚠ mismatch sim=0.81
  → google/gemini-3.1-flash-lite-preview     4 ( 0%) ⚠ mismatch sim=0.89
  → anthropic/claude-opus-4.7                3 ( 0%) ⚠ mismatch sim=0.63
  → openai/gpt-5.4                           3 ( 0%) ⚠ mismatch sim=1.00
  → anthropic/claude-haiku-4.5               2 ( 0%) ⚠ mismatch sim=0.62
  → anthropic/claude-fable-5                 2 ( 0%) ⚠ mismatch sim=0.65
  → openai/gpt-5-mini                        1 ( 0%) ⚠ mismatch sim=0.82
  → qwen/qwen3-max                           1 ( 0%) ⚠ mismatch sim=0.63
  → openai/gpt-4o                            1 ( 0%) ⚠ mismatch sim=0.81
  → google/gemini-3.5-flash-lite             1 ( 0%) ⚠ mismatch sim=0.72
  → deepseek/deepseek-v4-pro                 1 ( 0%) ⚠ mismatch sim=0.71
  → anthropic/claude-sonnet-4.5              1 ( 0%) ⚠ mismatch sim=0.72
  → openai/gpt-5                             1 ( 0%) ⚠ mismatch sim=0.80
  → qwen/qwen3-next-80b-a3b-instruct         1 ( 0%) ⚠ mismatch sim=0.64
  → anthropic/claude-sonnet-4.6              1 ( 0%) ⚠ mismatch sim=0.68

Capability score distribution (auxiliary)

  0-49     ██████████████░░░░░░░░░░░  527
  50-69    ██░░░░░░░░░░░░░░░░░░░░░░░   70
  70-89    █████████░░░░░░░░░░░░░░░░  317
  90-100   █░░░░░░░░░░░░░░░░░░░░░░░░   26

Host integrity ranking

Host	Runs	Average score	Family match	Most common V3 match
https://api.a6api.com/v1	34	22	0%	z-ai/glm-5.2 (3)
https://api5.orangeai.cc/v1	32	54	0%	z-ai/glm-5.2 (9)
https://llm.goaichat.top/v1	22	35	0%	—
https://api.x5m5x.com/v1	21	49	0%	z-ai/glm-5.2 (11)
https://api4.orangeai.cc/v1	21	42	0%	z-ai/glm-5.1 (1)
https://tokenrhythm.studio/v1	19	57	0%	z-ai/glm-5.2 (6)
https://api.gnrt.dev/v1	19	44	0%	—
https://yuanyuaicloud.cn/v1	19	45	0%	z-ai/glm-5.2 (1)
https://token.sensenova.cn/v1	16	33	0%	z-ai/glm-5.2 (2)
https://sharellm.cn/v1	15	28	0%	—
https://api.hcnsec.cn/v1	14	29	0%	z-ai/glm-5.1 (1)
https://router.lobakmerah.my.id/v1	12	67	0%	qwen/qwen3-max (1)
https://api.inferhub.dev/v1	11	14	0%	z-ai/glm-5 (2)
https://www.getunikey.ai/v1	10	11	0%	—
https://llm.xinlicloud.top/v1	10	36	0%	z-ai/glm-5 (2)
https://bandelbanget.xyz/v1	9	16	0%	z-ai/glm-5 (1)
https://a6api.com/v1	9	19	0%	z-ai/glm-5 (1)
https://seekai.cc/v1	9	17	0%	z-ai/glm-5.2 (1)
https://api.supxh.xin/v1	9	43	0%	—
https://api.sfkey.cn/v1	9	53	0%	z-ai/glm-5.1 (1)
https://ent.yzcld.com/v1	9	70	0%	—
https://speed.toter.me/v1	8	10	0%	—
https://ai.kaiplay.my.id/v1	8	82	0%	z-ai/glm-5.2 (2)
https://newapi.noo.cool/v1	8	7	0%	google/gemini-3.5-flash-lite (1)
https://ai.lnapi.cn/v1	8	80	0%	z-ai/glm-5.2 (4)

Recent anomalies

https://debug.tame.pwSep 12, 2026, 11:26 AM

Self-claimed model: glm-5.2
V2 predicted family: zhipu
V3 best match: —
Verdict: V3 low confidence · Abstain

https://9router.kliksosmed.id/v1Sep 11, 2026, 10:18 PM

Self-claimed model: prk/z-ai/glm-5.2
V2 predicted family: zhipu
V3 best match: —
Verdict: Family mismatch

https://www.getunikey.ai/v1Sep 11, 2026, 7:25 PM

Self-claimed model: z-ai/glm-5.2
V2 predicted family: zhipu
V3 best match: —
Verdict: Family mismatch

https://api.lmuai.com/v1Sep 11, 2026, 2:36 PM

Self-claimed model: glm-5.2
V2 predicted family: zhipu
V3 best match: z-ai/glm-5 (sim 0.61)
Verdict: Full swap

https://api.lmuai.comSep 11, 2026, 10:38 AM

Self-claimed model: glm-5.2
V2 predicted family: zhipu
V3 best match: —
Verdict: Full swap

https://9router.kliksosmed.id/v1Sep 10, 2026, 3:19 PM

Self-claimed model: ry/glm-5.2
V2 predicted family: zhipu
V3 best match: —
Verdict: V3 low confidence · Abstain

https://api.unreallou.com/v1Sep 10, 2026, 1:06 PM

Self-claimed model: glm-5.2
V2 predicted family: zhipu
V3 best match: —
Verdict: Family mismatch

https://zzone.cc.cd/v1Sep 10, 2026, 8:22 AM

Self-claimed model: glm-5.2
V2 predicted family: zhipu
V3 best match: —
Verdict: V3 low confidence · Abstain

http://103.250.11.160:28152/v1Sep 9, 2026, 3:19 PM

Self-claimed model: glm-5.2
V2 predicted family: zhipu
V3 best match: —
Verdict: V3 low confidence · Abstain

https://bandelbanget.xyz/v1Sep 9, 2026, 12:11 PM

Self-claimed model: glm-5.2
V2 predicted family: zhipu
V3 best match: —
Verdict: V3 low confidence · Abstain

https://dashscope-intl.aliyuncs.com/compatible-mode/v1Sep 9, 2026, 9:03 AM

Self-claimed model: glm-5.2
V2 predicted family: zhipu
V3 best match: —
Verdict: V3 low confidence · Abstain

https://dashscope-intl.aliyuncs.com/compatible-mode/v1Sep 9, 2026, 8:15 AM

Self-claimed model: glm-5.2
V2 predicted family: zhipu
V3 best match: —
Verdict: V3 low confidence · Abstain

https://token.sensenova.cn/v1Sep 9, 2026, 4:51 AM

Self-claimed model: glm-5.2
V2 predicted family: zhipu
V3 best match: —
Verdict: V3 low confidence · Abstain

https://a6api.com/v1Sep 8, 2026, 6:54 PM

Self-claimed model: glm-5.2
V2 predicted family: zhipu
V3 best match: z-ai/glm-5 (sim 0.61)
Verdict: Full swap

https://a6api.com/v1Sep 8, 2026, 6:46 PM

Self-claimed model: glm-5.2
V2 predicted family: zhipu
V3 best match: —
Verdict: V3 low confidence · Abstain

https://api.a6api.com/v1Sep 8, 2026, 4:16 PM

Self-claimed model: glm-5.2
V2 predicted family: zhipu
V3 best match: —
Verdict: V3 low confidence · Abstain

https://api.a6api.com/v1Sep 8, 2026, 4:02 PM

Self-claimed model: glm-5.2
V2 predicted family: zhipu
V3 best match: —
Verdict: Family mismatch

https://new-api.kunkunout.com/v1Sep 8, 2026, 4:02 AM

Self-claimed model: glm-5.2
V2 predicted family: zhipu
V3 best match: —
Verdict: Full swap

https://shiteru.id/v1Sep 7, 2026, 5:16 PM

Self-claimed model: glm-5.2
V2 predicted family: zhipu
V3 best match: —
Verdict: V3 low confidence · Abstain

https://according-conditions-self-prospects.trycloudflare.com/v1Sep 7, 2026, 2:14 PM

Self-claimed model: glm-5.2
V2 predicted family: zhipu
V3 best match: —
Verdict: V3 low confidence · Abstain

Methodology

BazaarLink Probe combines a Bayesian V2 family classifier across nine fingerprint categories with V3 sub-model references to verify endpoints claiming glm52. Low family consistency, a different V3 match, or divergence from selfClaim is flagged as anomalous. BazaarLink Probe home.

Test an endpoint yourself →