deepseek/deepseek-v4-pro · Model identity and integrity report
Last updated：Jul 30, 2026, 12:07 PM · Data window: past 90 days
Verification capability — fingerprint reference available
- Family
- deepseek
- 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
- 7 weighted rules · 4 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)
212 runs · 97 hosts · 103 anomalies
Family-fingerprint consistency
deepseek 138 ████████████████░░░░░░░░░ 65% (expected) unknown 29 ████░░░░░░░░░░░░░░░░░░░░░ 14% (suspicious) openai 26 ███░░░░░░░░░░░░░░░░░░░░░░ 12% (suspicious) anthropic 18 ██░░░░░░░░░░░░░░░░░░░░░░░ 8% (suspicious) google 1 ░░░░░░░░░░░░░░░░░░░░░░░░░ 0% (suspicious)
Observed V3 sub-model mapping
→ abstain 88 (42%) — low confidence → deepseek/deepseek-v4-pro 69 (33%) ✓ correct sim=0.79 → unknown 36 (17%) ⚠ mismatch → deepseek/deepseek-v4-flash 9 ( 4%) ⚠ mismatch sim=0.79 → deepseek/deepseek-chat-v3.1 7 ( 3%) ⚠ mismatch sim=0.76 → anthropic/claude-haiku-4.5 2 ( 1%) ⚠ mismatch sim=0.62 → google/gemini-2.5-flash-lite 1 ( 0%) ⚠ mismatch sim=0.71
Capability score distribution (auxiliary)
0-49 ██████████████░░░░░░░░░░░ 120 50-69 ██░░░░░░░░░░░░░░░░░░░░░░░ 19 70-89 ████████░░░░░░░░░░░░░░░░░ 63 90-100 █░░░░░░░░░░░░░░░░░░░░░░░░ 10
Host integrity ranking
| Host | Runs | Average score | Family match | Most common V3 match |
|---|---|---|---|---|
| https://api.deepseek.com/v1 | 41 | 28 | 46% | deepseek/deepseek-v4-pro (17) |
| https://api.183399.xyz/v1 | 9 | 19 | 100% | deepseek/deepseek-chat-v3.1 (2) |
| https://api.hcnsec.cn/v1 | 8 | 46 | 38% | deepseek/deepseek-chat-v3.1 (1) |
| https://card.nassaapi.xyz/v1 | 8 | 38 | 63% | deepseek/deepseek-v4-flash (3) |
| https://api.yuk15n0w.asia/v1 | 6 | 59 | 100% | deepseek/deepseek-v4-flash (1) |
| https://tp-api.chinadatapay.com:8000/v1 | 4 | 83 | 100% | deepseek/deepseek-v4-pro (4) |
| https://tian-shu.org/v1 | 4 | 61 | 75% | deepseek/deepseek-v4-pro (3) |
| https://52mx.net/v1 | 4 | 48 | 75% | — |
| https://api.a6api.com/v1 | 3 | 82 | 100% | deepseek/deepseek-v4-pro (1) |
| https://a6api.com/v1 | 3 | 28 | 100% | deepseek/deepseek-v4-pro (1) |
| https://api.x5m5x.com/v1 | 3 | 27 | 100% | deepseek/deepseek-chat-v3.1 (1) |
| https://api.nassaapi.xyz/v1 | 3 | 72 | 100% | deepseek/deepseek-v4-flash (2) |
| https://api.kukuit.com/v1 | 3 | 21 | 33% | deepseek/deepseek-v4-pro (1) |
| https://api.tokengo.com/v1 | 3 | 3 | 33% | deepseek/deepseek-v4-pro (1) |
| https://luna-ai.shop/v1 | 3 | 87 | 67% | deepseek/deepseek-v4-pro (2) |
| https://ai.pianyiapi.com/v1 | 3 | 0 | 33% | — |
| https://api.iamhc.cn/v1 | 3 | 0 | 0% | — |
| https://www.xinsuanai.com/v1 | 3 | 0 | 0% | — |
| https://taotoken.net/api/v1 | 2 | 25 | 0% | — |
| https://api5.orangeai.cc/v1 | 2 | 92 | 100% | deepseek/deepseek-chat-v3.1 (1) |
| https://api.aitunnel.ru/v1 | 2 | 0 | 100% | deepseek/deepseek-v4-pro (2) |
| https://web-ai-media-editor.cn/v1 | 2 | 27 | 100% | deepseek/deepseek-v4-flash (2) |
| https://inferaichat.com/v1 | 2 | 21 | 50% | deepseek/deepseek-chat-v3.1 (1) |
| https://api.vectorengine.ai/v1 | 2 | 47 | 100% | deepseek/deepseek-v4-pro (2) |
| https://openapi.cloud-ai.cn/v1 | 2 | 38 | 50% | deepseek/deepseek-v4-pro (1) |
Recent anomalies
https://ai.furry.vg/v1Jul 30, 2026, 11:44 AM
- Self-claimed model
- deepseek-v4-pro
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://api.yuk15n0w.asia/v1Jul 30, 2026, 8:55 AM
- Self-claimed model
- deepseek-v4-pro
- V2 predicted family
- deepseek
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://api.yuk15n0w.asia/v1Jul 30, 2026, 8:50 AM
- Self-claimed model
- deepseek-v4-pro
- V2 predicted family
- deepseek
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://api.xkiro.com/v1Jul 30, 2026, 4:07 AM
- Self-claimed model
- deepseek/deepseek-v4-pro
- V2 predicted family
- deepseek
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://llm.xinlicloud.top/v1Jul 29, 2026, 6:01 AM
- Self-claimed model
- deepseek-v4-pro
- V2 predicted family
- anthropic
- V3 best match
- —
- Verdict
- Full swap
https://www.dxnt.com/v1Jul 28, 2026, 8:44 AM
- Self-claimed model
- deepseek-v4-pro
- V2 predicted family
- deepseek
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://api5.orangeai.cc/v1Jul 27, 2026, 7:12 AM
- Self-claimed model
- deepseek-v4-pro
- V2 predicted family
- deepseek
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://api.sta1n.site/v1Jul 27, 2026, 1:22 AM
- Self-claimed model
- [免费]deepseek-v4-pro
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- Family mismatch
https://api.a6api.com/v1Jul 26, 2026, 1:56 PM
- Self-claimed model
- deepseek-v4-pro
- V2 predicted family
- deepseek
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://othersapi.com/v1Jul 26, 2026, 6:10 AM
- Self-claimed model
- deepseek-v4-pro
- V2 predicted family
- deepseek
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://api.yuk15n0w.asia/v1Jul 26, 2026, 4:47 AM
- Self-claimed model
- deepseek-v4-pro
- V2 predicted family
- deepseek
- V3 best match
- deepseek/deepseek-v4-flash (sim 0.67)
- Verdict
- Sub-model downgrade
https://web-ai-media-editor.cn/v1Jul 26, 2026, 1:32 AM
- Self-claimed model
- deepseek-v4-pro
- V2 predicted family
- deepseek
- V3 best match
- deepseek/deepseek-v4-flash (sim 0.85)
- Verdict
- Sub-model downgrade
https://inferaichat.com/v1Jul 26, 2026, 1:23 AM
- Self-claimed model
- deepseek-v4-pro
- V2 predicted family
- deepseek
- V3 best match
- deepseek/deepseek-chat-v3.1 (sim 0.76)
- Verdict
- Full swap
https://web-ai-media-editor.cn/v1Jul 26, 2026, 1:11 AM
- Self-claimed model
- deepseek-v4-pro
- V2 predicted family
- deepseek
- V3 best match
- deepseek/deepseek-v4-flash (sim 0.65)
- Verdict
- Sub-model downgrade
https://api.3mc.net/v1Jul 25, 2026, 7:11 PM
- Self-claimed model
- DeepSeek-V4-Pro
- V2 predicted family
- deepseek
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://api.deepseek.com/v1Jul 25, 2026, 3:55 PM
- Self-claimed model
- deepseek-v4-pro
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://api.hcnsec.cn/v1Jul 25, 2026, 9:21 AM
- Self-claimed model
- DeepSeek-V4-Pro
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://a6api.com/v1Jul 25, 2026, 8:13 AM
- Self-claimed model
- deepseek-v4-pro
- V2 predicted family
- deepseek
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://api.hcnsec.cn/v1Jul 24, 2026, 8:56 PM
- Self-claimed model
- DeepSeek-V4-Pro
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://api.183399.xyz/v1Jul 24, 2026, 5:00 PM
- Self-claimed model
- deepseek-v4-pro
- V2 predicted family
- deepseek
- V3 best match
- deepseek/deepseek-chat-v3.1 (sim 0.66)
- Verdict
- Sub-model downgrade
Methodology
BazaarLink Probe combines a Bayesian V2 family classifier across nine fingerprint categories with V3 sub-model references to verify endpoints claiming deepseek/deepseek-v4-pro. Low family consistency, a different V3 match, or divergence from selfClaim is flagged as anomalous. BazaarLink Probe home.