← Back to model

gpt52 · Model identity and integrity report

Last updated：Sep 13, 2026, 4:24 PM · Data window: past 90 days

Verification capability — fingerprint reference available

Family: openai
V3 sub-model reference: Available ✓
Checks: 知識截止 · 能力探針 · 拒絕模板 · 拒絕溫度參數
Family-fingerprint coverage: 52 weighted rules · 8 feature categories

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

Measured results (90 days)

12 runs · 8 hosts · 5 anomalies

Family-fingerprint consistency

  openai           5  ███████████░░░░░░░░░░░░░░   42% (expected)
  unknown          4  ████████░░░░░░░░░░░░░░░░░   33% (suspicious)
  anthropic        2  ████░░░░░░░░░░░░░░░░░░░░░   17% (suspicious)
  qwen             1  ██░░░░░░░░░░░░░░░░░░░░░░░    8% (suspicious)

Observed V3 sub-model mapping

  → abstain                                  5 (42%) — low confidence
  → unknown                                  4 (33%) ⚠ mismatch
  → openai/gpt-5.2                           3 (25%) ⚠ mismatch sim=0.84

Capability score distribution (auxiliary)

  0-49     ███████████████████░░░░░░    9
  50-69    ██░░░░░░░░░░░░░░░░░░░░░░░    1
  70-89    ██░░░░░░░░░░░░░░░░░░░░░░░    1
  90-100   ██░░░░░░░░░░░░░░░░░░░░░░░    1

Host integrity ranking

Host	Runs	Average score	Family match	Most common V3 match
https://api.a6api.com/v1	2	43	100%	openai/gpt-5.2 (2)
https://llm-api.net/v1	2	0	50%	—
https://codex.hiyo.top/v1	2	0	0%	—
https://api.quotaflow.ai/openai/v1	2	72	50%	openai/gpt-5.2 (1)
https://ai.meiaodai.xyz/v1	1	0	0%	—
https://vip.j3gb.com/v1	1	0	100%	—
https://omaleai.qzz.io/v1	1	39	0%	—
http://api.halphen.cn/openai/v1	1	0	0%	—

Recent anomalies

https://ai.meiaodai.xyz/v1Aug 22, 2026, 9:37 PM

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

https://llm-api.net/v1Jul 22, 2026, 2:33 PM

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

https://llm-api.net/v1Jul 22, 2026, 2:11 PM

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

https://vip.j3gb.com/v1Jul 9, 2026, 11:23 AM

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

http://api.halphen.cn/openai/v1Jun 16, 2026, 3:51 AM

Self-claimed model: gpt-5.2
V2 predicted family: anthropic
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 gpt52. Low family consistency, a different V3 match, or divergence from selfClaim is flagged as anomalous. BazaarLink Probe home.

Test an endpoint yourself →