claudeopus5 · Model identity and integrity report

289 runs · 141 hosts · 278 anomalies

Family-fingerprint consistency

  anthropic      276  ████████████████████████░   96% (suspicious)
  unknown         11  █░░░░░░░░░░░░░░░░░░░░░░░░    4% (suspicious)
  openai           2  ░░░░░░░░░░░░░░░░░░░░░░░░░    1% (suspicious)

Observed V3 sub-model mapping

  → abstain                                125 (43%) — low confidence
  → anthropic/claude-opus-4.7               60 (21%) ⚠ mismatch sim=0.80
  → anthropic/claude-opus-4.5               47 (16%) ⚠ mismatch sim=0.68
  → unknown                                 25 ( 9%) ⚠ mismatch
  → anthropic/claude-sonnet-4.6             12 ( 4%) ⚠ mismatch sim=0.72
  → anthropic/claude-opus-4.6                9 ( 3%) ⚠ mismatch sim=0.83
  → anthropic/claude-fable-5                 8 ( 3%) ⚠ mismatch sim=0.97
  → anthropic/claude-sonnet-4.5              3 ( 1%) ⚠ mismatch sim=0.62

Capability score distribution (auxiliary)

  0-49     ████████████░░░░░░░░░░░░░  135
  50-69    █░░░░░░░░░░░░░░░░░░░░░░░░   12
  70-89    ███████░░░░░░░░░░░░░░░░░░   75
  90-100   ██████░░░░░░░░░░░░░░░░░░░   67

BazaarLink Probe combines a Bayesian V2 family classifier across nine fingerprint categories with V3 sub-model references to verify endpoints claiming claudeopus5. Low family consistency, a different V3 match, or divergence from selfClaim is flagged as anomalous. BazaarLink Probe home.