openai/gpt-5.3-codex · Model identity and integrity report
Last updated：Sep 13, 2026, 7:44 PM · Data window: past 90 days
Measured results (90 days)
11 runs · 8 hosts · 5 anomalies
Family-fingerprint consistency
unknown 6 ██████████████░░░░░░░░░░░ 55% (suspicious) openai 5 ███████████░░░░░░░░░░░░░░ 45% (suspicious)
Observed V3 sub-model mapping
→ unknown 6 (55%) ⚠ mismatch → abstain 4 (36%) — low confidence → openai/gpt-4.1 1 ( 9%) ⚠ mismatch sim=0.80
Capability score distribution (auxiliary)
0-49 ████████████████░░░░░░░░░ 7 50-69 ░░░░░░░░░░░░░░░░░░░░░░░░░ 0 70-89 ███████░░░░░░░░░░░░░░░░░░ 3 90-100 ██░░░░░░░░░░░░░░░░░░░░░░░ 1
Host integrity ranking
| Host | Runs | Average score | Family match | Most common V3 match |
|---|---|---|---|---|
| http://187.53.135.158:7101/v1 | 3 | 0 | 0% | — |
| https://api.quotaflow.ai/openai/v1 | 2 | 51 | 0% | — |
| https://llm.twinight.co/v1 | 1 | 0 | 0% | — |
| https://jizhiapi.site/v1 | 1 | 0 | 0% | — |
| https://api.bluesminds.com/v1 | 1 | 79 | 0% | — |
| https://sarang.id/v1 | 1 | 88 | 0% | openai/gpt-4.1 (1) |
| https://tokenapi.link/v1 | 1 | 88 | 0% | — |
| https://ciyuan.fast/v1 | 1 | 0 | 0% | — |
Recent anomalies
https://llm.twinight.co/v1Sep 6, 2026, 11:36 AM
- Self-claimed model
- gpt-5.3-codex
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://api.bluesminds.com/v1Aug 25, 2026, 3:49 PM
- Self-claimed model
- gpt-5.3-codex
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://sarang.id/v1Jul 15, 2026, 12:40 PM
- Self-claimed model
- gpt-5.3-codex
- V2 predicted family
- openai
- V3 best match
- openai/gpt-4.1 (sim 0.80)
- Verdict
- Full swap
https://tokenapi.link/v1Jul 15, 2026, 3:48 AM
- Self-claimed model
- gpt-5.3-codex
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://api.quotaflow.ai/openai/v1Jun 22, 2026, 6:42 AM
- Self-claimed model
- gpt-5.3-codex
- 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 openai/gpt-5.3-codex. Low family consistency, a different V3 match, or divergence from selfClaim is flagged as anomalous. BazaarLink Probe home.