kimik3 · Model identity and integrity report
Last updated:Sep 13, 2026, 8:05 PM · Data window: past 90 days
Measured results (90 days)
574 runs · 237 hosts · 499 anomalies
Family-fingerprint consistency
anthropic 320 ██████████████░░░░░░░░░░░ 56% (suspicious) openai 161 ███████░░░░░░░░░░░░░░░░░░ 28% (suspicious) unknown 75 ███░░░░░░░░░░░░░░░░░░░░░░ 13% (suspicious) qwen 7 ░░░░░░░░░░░░░░░░░░░░░░░░░ 1% (suspicious) deepseek 6 ░░░░░░░░░░░░░░░░░░░░░░░░░ 1% (suspicious) zhipu 3 ░░░░░░░░░░░░░░░░░░░░░░░░░ 1% (suspicious) google 2 ░░░░░░░░░░░░░░░░░░░░░░░░░ 0% (suspicious)
Observed V3 sub-model mapping
→ abstain 411 (72%) — low confidence → unknown 115 (20%) ⚠ mismatch → anthropic/claude-sonnet-5 21 ( 4%) ⚠ mismatch sim=0.74 → anthropic/claude-sonnet-4.5 7 ( 1%) ⚠ mismatch sim=0.82 → anthropic/claude-opus-4.7 6 ( 1%) ⚠ mismatch sim=0.78 → openai/gpt-5 5 ( 1%) ⚠ mismatch sim=0.84 → anthropic/claude-opus-4.5 4 ( 1%) ⚠ mismatch sim=0.74 → anthropic/claude-fable-5 2 ( 0%) ⚠ mismatch sim=0.90 → deepseek/deepseek-v4-pro 1 ( 0%) ⚠ mismatch sim=0.64 → openai/gpt-5-mini 1 ( 0%) ⚠ mismatch sim=0.82 → openai/gpt-5.4 1 ( 0%) ⚠ mismatch sim=0.81
Capability score distribution (auxiliary)
0-49 ███████████████████░░░░░░ 438 50-69 ██░░░░░░░░░░░░░░░░░░░░░░░ 54 70-89 ████░░░░░░░░░░░░░░░░░░░░░ 82 90-100 ░░░░░░░░░░░░░░░░░░░░░░░░░ 0
Host integrity ranking
| Host | Runs | Average score | Family match | Most common V3 match |
|---|---|---|---|---|
| https://api.a6api.com/v1 | 68 | 9 | 0% | anthropic/claude-sonnet-4.5 (3) |
| https://api.gnrt.dev/v1 | 30 | 26 | 0% | — |
| https://bansos.hshino.dev/v1 | 21 | 12 | 0% | openai/gpt-5 (1) |
| https://api.hcnsec.cn/v1 | 12 | 16 | 0% | — |
| https://bandelbanget.xyz/v1 | 12 | 7 | 0% | openai/gpt-5-mini (1) |
| https://api.inferhub.dev/v1 | 12 | 11 | 0% | anthropic/claude-opus-4.5 (1) |
| https://madefaka.my.id/v1 | 12 | 12 | 0% | — |
| https://llm.goaichat.top/v1 | 11 | 7 | 0% | — |
| https://ai.redromance.pro/v1 | 9 | 35 | 0% | — |
| https://speed.toter.me/v1 | 7 | 31 | 0% | anthropic/claude-opus-4.5 (1) |
| https://api.blackbox.ai/v1 | 7 | 61 | 0% | anthropic/claude-sonnet-5 (4) |
| https://api.amanai.dev/v1 | 7 | 0 | 0% | anthropic/claude-sonnet-5 (2) |
| https://api.moonshot.cn/v1 | 6 | 20 | 0% | anthropic/claude-opus-4.7 (1) |
| https://www.getunikey.ai/v1 | 6 | 12 | 0% | — |
| https://sharellm.cn/v1 | 6 | 36 | 0% | — |
| https://www.sharellm.cn/v1 | 6 | 0 | 0% | — |
| https://ai.kaiplay.my.id/v1 | 6 | 53 | 0% | — |
| https://www.canwaycloud.ai/v1 | 6 | 0 | 0% | — |
| https://a6api.com/v1 | 5 | 0 | 0% | — |
| https://ai.furry.vg/v1 | 5 | 20 | 0% | — |
| https://gregy.nbhooks.me/v1 | 5 | 42 | 0% | — |
| https://api.supxh.xin/v1 | 5 | 16 | 0% | — |
| https://ai.baipiao.co/v1 | 4 | 0 | 0% | — |
| https://router.kiosapi.com/v1 | 4 | 0 | 0% | openai/gpt-5 (2) |
| https://api.52token.org/v1 | 4 | 0 | 0% | — |
Recent anomalies
https://api.a6api.com/v1Sep 13, 2026, 1:33 PM
- Self-claimed model
- kimi-k3
- V2 predicted family
- anthropic
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://pianyitoken.gay/v1Sep 13, 2026, 1:01 PM
- Self-claimed model
- 订阅-kimi-k3
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://api.vleee.net/v1Sep 13, 2026, 11:52 AM
- Self-claimed model
- cb/kimi-k3
- V2 predicted family
- anthropic
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://utopiapps.online/v1Sep 13, 2026, 7:43 AM
- Self-claimed model
- cbai/kimi-k3
- V2 predicted family
- anthropic
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://api.a6api.com/v1Sep 13, 2026, 7:18 AM
- Self-claimed model
- kimi-k3
- V2 predicted family
- anthropic
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://ces.ddddai.xyz/v1Sep 13, 2026, 6:48 AM
- Self-claimed model
- kimi-k3
- V2 predicted family
- deepseek
- V3 best match
- deepseek/deepseek-v4-pro (sim 0.64)
- Verdict
- Full swap
https://api.a6api.com/v1Sep 13, 2026, 6:33 AM
- Self-claimed model
- kimi-k3
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://deepseek.yunjunet.cn/v1Sep 13, 2026, 5:41 AM
- Self-claimed model
- kimi-k3
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://a6api.com/v1Sep 12, 2026, 8:11 PM
- Self-claimed model
- kimi-k3
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://a6api.com/v1Sep 12, 2026, 8:07 PM
- Self-claimed model
- kimi-k3
- V2 predicted family
- anthropic
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://api.nbhooks.my.id/v1Sep 12, 2026, 1:33 PM
- Self-claimed model
- Greg/kimi-k3
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://api.a6api.com/v1Sep 12, 2026, 1:24 PM
- Self-claimed model
- kimi-k3
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://cf.hetune.top/v1Sep 12, 2026, 11:21 AM
- Self-claimed model
- kimi-k3
- V2 predicted family
- anthropic
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://othersapi.com/v1Sep 12, 2026, 10:55 AM
- Self-claimed model
- kimi-k3
- V2 predicted family
- anthropic
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://token.firstpy.com/v1Sep 12, 2026, 6:43 AM
- Self-claimed model
- kimi-k3
- V2 predicted family
- anthropic
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://llm.goaichat.top/v1Sep 12, 2026, 12:53 AM
- Self-claimed model
- kimi-k3
- V2 predicted family
- anthropic
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://llm.goaichat.top/v1Sep 12, 2026, 12:47 AM
- Self-claimed model
- kimi-k3
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://www.micuapi.ai/v1Sep 12, 2026, 12:05 AM
- Self-claimed model
- kimi-k3
- V2 predicted family
- anthropic
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://api.a6api.com/v1Sep 11, 2026, 10:14 PM
- Self-claimed model
- kimi-k3
- V2 predicted family
- openai
- V3 best match
- —
- Verdict
- V3 low confidence · Abstain
https://api.a6api.com/v1Sep 11, 2026, 7:31 PM
- Self-claimed model
- kimi-k3
- 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 kimik3. Low family consistency, a different V3 match, or divergence from selfClaim is flagged as anomalous. BazaarLink Probe home.