gpt54 · Model identity and integrity report
Last updated：Sep 13, 2026, 8:53 PM · Data window: past 90 days
Measured results (90 days)
320 runs · 202 hosts · 295 anomalies
Family-fingerprint consistency
openai 290 ███████████████████████░░ 91% (suspicious) unknown 25 ██░░░░░░░░░░░░░░░░░░░░░░░ 8% (suspicious) anthropic 2 ░░░░░░░░░░░░░░░░░░░░░░░░░ 1% (suspicious) deepseek 2 ░░░░░░░░░░░░░░░░░░░░░░░░░ 1% (suspicious) zhipu 1 ░░░░░░░░░░░░░░░░░░░░░░░░░ 0% (suspicious)
Observed V3 sub-model mapping
→ openai/gpt-5.4 156 (49%) ⚠ mismatch sim=0.98 → abstain 118 (37%) — low confidence → unknown 34 (11%) ⚠ mismatch → openai/gpt-5.4-mini 7 ( 2%) ⚠ mismatch sim=0.96 → openai/gpt-5-mini 2 ( 1%) ⚠ mismatch sim=0.78 → deepseek/deepseek-v4-pro 1 ( 0%) ⚠ mismatch sim=0.97 → deepseek/deepseek-v4-flash 1 ( 0%) ⚠ mismatch sim=0.84 → openai/gpt-5.5 1 ( 0%) ⚠ mismatch sim=0.79
Capability score distribution (auxiliary)
0-49 ███████████████░░░░░░░░░░ 189 50-69 ░░░░░░░░░░░░░░░░░░░░░░░░░ 4 70-89 ██████░░░░░░░░░░░░░░░░░░░ 80 90-100 ████░░░░░░░░░░░░░░░░░░░░░ 47
Host integrity ranking
| Host | Runs | Average score | Family match | Most common V3 match |
|---|---|---|---|---|
| https://uf.raoqiuqiu.fun/v1 | 14 | 0 | 0% | — |
| https://aihub.top/v1 | 9 | 0 | 0% | — |
| https://jojocode.com/v1 | 7 | 0 | 0% | openai/gpt-5.4 (6) |
| https://ai.input.im/v1 | 7 | 63 | 0% | openai/gpt-5.4 (3) |
| http://187.53.135.158:7101/v1 | 5 | 0 | 0% | — |
| https://tokenapi.link/v1 | 5 | 34 | 0% | openai/gpt-5.4 (4) |
| https://ruoli.dev/v1 | 4 | 0 | 0% | openai/gpt-5.4 (1) |
| https://zyrus.aitoken.credit/v1 | 4 | 92 | 0% | openai/gpt-5.4 (2) |
| https://cf-fast.cctq.ai/v1 | 4 | 44 | 0% | openai/gpt-5.4 (3) |
| https://a6api.com/v1 | 4 | 92 | 0% | openai/gpt-5.4 (3) |
| https://aicanapi.com/v1 | 4 | 45 | 0% | openai/gpt-5.4 (2) |
| https://ai.furry.vg/v1 | 4 | 85 | 0% | openai/gpt-5.4 (3) |
| https://api.a6api.com/v1 | 3 | 29 | 0% | openai/gpt-5.4 (1) |
| https://www.micuapi.ai/v1 | 3 | 60 | 0% | openai/gpt-5.4 (3) |
| https://ai-api.kkidc.com/v1 | 3 | 31 | 0% | openai/gpt-5.4 (2) |
| https://api.aiapipay.com/v1 | 3 | 89 | 0% | openai/gpt-5.4 (2) |
| https://botcf.com/v1 | 3 | 61 | 0% | openai/gpt-5.4 (3) |
| https://api.wluvyh.cloud/v1 | 3 | 50 | 0% | openai/gpt-5.4 (1) |
| https://ciyuan.fast/v1 | 3 | 0 | 0% | — |
| https://muyuan.do/v1 | 3 | 36 | 0% | openai/gpt-5.4-mini (1) |
| https://smz6.com/v1 | 3 | 0 | 0% | openai/gpt-5.4 (3) |
| https://www.api-codex.cn/v1 | 3 | 27 | 0% | — |
| https://api.vectorengine.ai/v1 | 3 | 57 | 0% | openai/gpt-5.4 (2) |
| https://new.lingj.online/v1 | 2 | 83 | 0% | — |
| https://api.kgapi.net/v1 | 2 | 0 | 0% | — |
Recent anomalies
https://uf.raoqiuqiu.fun/v1Sep 11, 2026, 3:00 AM
- Self-claimed model
- gpt-5.4
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://uf.raoqiuqiu.fun/v1Sep 10, 2026, 6:31 AM
- Self-claimed model
- gpt-5.4
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://api.soloapi.cc/v1Sep 5, 2026, 10:37 AM
- Self-claimed model
- gpt-5.4
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://new.lingj.online/v1Sep 5, 2026, 12:27 AM
- Self-claimed model
- gpt-5.4
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://new.lingj.online/v1Sep 5, 2026, 12:24 AM
- Self-claimed model
- gpt-5.4
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://sharellm.cn/v1Sep 3, 2026, 9:18 AM
- Self-claimed model
- gpt-5.4
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://new.fluapi.com/v1Aug 31, 2026, 8:27 PM
- Self-claimed model
- gpt-5.4
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://www.tokenkey.uk/v1Aug 30, 2026, 7:17 PM
- Self-claimed model
- gpt-5.4
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://quickaicode.com/v1Aug 30, 2026, 2:12 PM
- Self-claimed model
- gpt-5.4
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://www.maitokens.top/v1Aug 30, 2026, 1:52 PM
- Self-claimed model
- gpt-5.4
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://uf.raoqiuqiu.fun/v1Aug 30, 2026, 1:12 PM
- Self-claimed model
- gpt-5.4
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://uf.raoqiuqiu.fun/v1Aug 30, 2026, 1:10 PM
- Self-claimed model
- gpt-5.4
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://uf.raoqiuqiu.fun/v1Aug 30, 2026, 3:27 AM
- Self-claimed model
- gpt-5.4
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://ezaiapi.com/v1Aug 28, 2026, 10:54 AM
- Self-claimed model
- gpt-5.4
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- Family mismatch
https://api.a6api.com/v1Aug 27, 2026, 1:34 PM
- Self-claimed model
- gpt-5.4
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- Family mismatch
https://api.modelshare.dev/v1Aug 26, 2026, 1:26 PM
- Self-claimed model
- gpt-5.4
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://seekai.cc/v1Aug 26, 2026, 5:17 AM
- Self-claimed model
- gpt-5-4
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://farhanhana.my.id/v1Aug 25, 2026, 5:50 AM
- Self-claimed model
- gpt-5.4
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://xb1520.com/v1Aug 24, 2026, 5:16 PM
- Self-claimed model
- gpt-5.4
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://aihub.top/v1Aug 24, 2026, 5:09 PM
- Self-claimed model
- gpt-5.4
- V2 predicted family
- openai
- 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 gpt54. Low family consistency, a different V3 match, or divergence from selfClaim is flagged as anomalous. BazaarLink Probe home.