google/gemini-4-argon · Model identity and integrity report

Family

google

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

17 weighted rules · 5 feature categories

Concrete reference answers, refusal templates, and individual signal weights are not published because disclosure would help spoofing endpoints evade detection.

128 runs · 66 hosts · 118 anomalies

Family-fingerprint consistency

  google          94  ██████████████████░░░░░░░   73% (expected)
  anthropic       18  ████░░░░░░░░░░░░░░░░░░░░░   14% (suspicious)
  openai           9  ██░░░░░░░░░░░░░░░░░░░░░░░    7% (suspicious)
  unknown          5  █░░░░░░░░░░░░░░░░░░░░░░░░    4% (suspicious)
  deepseek         2  █░░░░░░░░░░░░░░░░░░░░░░░░    2% (suspicious)

Observed V3 sub-model mapping

  → abstain                                 68 (53%) — low confidence
  → google/gemini-3.6-flash                 39 (30%) ⚠ mismatch sim=0.73
  → unknown                                 11 ( 9%) ⚠ mismatch
  → google/gemini-3.5-flash-lite             8 ( 6%) ⚠ mismatch sim=0.73
  → anthropic/claude-sonnet-5.5              1 ( 1%) ⚠ mismatch sim=0.82
  → google/gemini-3.1-flash-lite-preview     1 ( 1%) ⚠ mismatch sim=0.65

Capability score distribution (auxiliary)

  0-49     █████████████████████████  128
  50-69    ░░░░░░░░░░░░░░░░░░░░░░░░░    0
  70-89    ░░░░░░░░░░░░░░░░░░░░░░░░░    0
  90-100   ░░░░░░░░░░░░░░░░░░░░░░░░░    0

BazaarLink Probe combines a Bayesian V2 family classifier across nine fingerprint categories with V3 sub-model references to verify endpoints claiming google/gemini-4-argon. Low family consistency, a different V3 match, or divergence from selfClaim is flagged as anomalous. BazaarLink Probe home.