qwen37max · Model identity and integrity report
Last updated：Sep 13, 2026, 5:01 PM · Data window: past 90 days
Measured results (90 days)
19 runs · 17 hosts · 13 anomalies
Family-fingerprint consistency
qwen 12 ████████████████░░░░░░░░░ 63% (expected) openai 3 ████░░░░░░░░░░░░░░░░░░░░░ 16% (suspicious) anthropic 2 ███░░░░░░░░░░░░░░░░░░░░░░ 11% (suspicious) unknown 2 ███░░░░░░░░░░░░░░░░░░░░░░ 11% (suspicious)
Observed V3 sub-model mapping
→ abstain 15 (79%) — low confidence → unknown 3 (16%) ⚠ mismatch → openai/gpt-5-mini 1 ( 5%) ⚠ mismatch sim=0.82
Capability score distribution (auxiliary)
0-49 ████████████████░░░░░░░░░ 12 50-69 █░░░░░░░░░░░░░░░░░░░░░░░░ 1 70-89 ████████░░░░░░░░░░░░░░░░░ 6 90-100 ░░░░░░░░░░░░░░░░░░░░░░░░░ 0
Host integrity ranking
| Host | Runs | Average score | Family match | Most common V3 match |
|---|---|---|---|---|
| https://nut-george-qualify-bill.trycloudflare.com/v1 | 2 | 36 | 0% | — |
| https://api.relaycore.cn/v1 | 2 | 77 | 100% | — |
| https://api.ai.rhikaku.net/v1 | 1 | 0 | 100% | — |
| https://www.getunikey.ai/v1 | 1 | 73 | 100% | — |
| https://dashscope-intl.aliyuncs.com/compatible-mode/v1 | 1 | 0 | 100% | — |
| https://beizhi.sylu.cc/v1 | 1 | 57 | 0% | openai/gpt-5-mini (1) |
| https://api.xkiro.com/v1 | 1 | 0 | 100% | — |
| https://router.lobakmerah.my.id/v1 | 1 | 74 | 100% | — |
| https://api.routescope.ai/v1 | 1 | 0 | 100% | — |
| https://ai.kaiplay.my.id/v1 | 1 | 78 | 100% | — |
| https://kjapi.botsmart.net/v1 | 1 | 0 | 100% | — |
| https://66ai.ai/v1 | 1 | 0 | 0% | — |
| https://llm.xinlicloud.top/v1 | 1 | 42 | 0% | — |
| https://aibz.cc/v1 | 1 | 0 | 0% | — |
| https://newapi.xlsx.vip/v1 | 1 | 0 | 100% | — |
| https://ai.shiteru.id/v1 | 1 | 0 | 100% | — |
| https://api.vsllm.com/v1 | 1 | 0 | 0% | — |
Recent anomalies
https://www.getunikey.ai/v1Sep 4, 2026, 5:37 AM
- Self-claimed model
- qwen3.7-max
- V2 predicted family
- qwen
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://dashscope-intl.aliyuncs.com/compatible-mode/v1Aug 26, 2026, 4:41 PM
- Self-claimed model
- qwen3.7-max
- V2 predicted family
- qwen
- V3 best match
- —
- Verdict
- Full swap
https://beizhi.sylu.cc/v1Aug 24, 2026, 9:13 PM
- Self-claimed model
- qwen3.7-max
- V2 predicted family
- openai
- V3 best match
- openai/gpt-5-mini (sim 0.82)
- Verdict
- Full swap
https://nut-george-qualify-bill.trycloudflare.com/v1Aug 18, 2026, 3:27 AM
- Self-claimed model
- qd/qwen3.7-max
- V2 predicted family
- anthropic
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://nut-george-qualify-bill.trycloudflare.com/v1Aug 18, 2026, 3:23 AM
- Self-claimed model
- qd/qwen3.7-max
- V2 predicted family
- anthropic
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://api.xkiro.com/v1Aug 17, 2026, 12:22 PM
- Self-claimed model
- qwen/qwen3.7-max
- V2 predicted family
- qwen
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://api.routescope.ai/v1Aug 11, 2026, 2:07 AM
- Self-claimed model
- qwen3.7-max
- V2 predicted family
- qwen
- V3 best match
- —
- Verdict
- Full swap
https://api.relaycore.cn/v1Aug 7, 2026, 3:29 AM
- Self-claimed model
- qwen3.7-max
- V2 predicted family
- qwen
- V3 best match
- —
- Verdict
- Full swap
https://api.relaycore.cn/v1Aug 7, 2026, 3:21 AM
- Self-claimed model
- qwen3.7-max
- V2 predicted family
- qwen
- V3 best match
- —
- Verdict
- Full swap
https://kjapi.botsmart.net/v1Aug 5, 2026, 10:30 AM
- Self-claimed model
- qwen3.7-max-2026-06-08
- V2 predicted family
- qwen
- V3 best match
- —
- Verdict
- Full swap
https://llm.xinlicloud.top/v1Jul 29, 2026, 6:07 AM
- Self-claimed model
- qwen3.7-max
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://aibz.cc/v1Jul 22, 2026, 11:20 PM
- Self-claimed model
- qwen3.7-max
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://newapi.xlsx.vip/v1Jul 20, 2026, 10:05 AM
- Self-claimed model
- qwen-3.7-max
- V2 predicted family
- qwen
- V3 best match
- —
- Verdict
- Full swap
Methodology
BazaarLink Probe combines a Bayesian V2 family classifier across nine fingerprint categories with V3 sub-model references to verify endpoints claiming qwen37max. Low family consistency, a different V3 match, or divergence from selfClaim is flagged as anomalous. BazaarLink Probe home.