← Back to model

ultimate · Model identity and integrity report

Name: ultimate relay integrity probe aggregate
Creator: BazaarLink

Last updated：Aug 18, 2026, 5:28 AM · Data window: past 90 days

Measured results (90 days)

6 runs · 4 hosts · 6 anomalies

Family-fingerprint consistency

  anthropic        5  █████████████████████░░░░   83% (suspicious)
  qwen             1  ████░░░░░░░░░░░░░░░░░░░░░   17% (suspicious)

Observed V3 sub-model mapping

  → abstain                                  5 (83%) — low confidence
  → anthropic/claude-opus-4.7                1 (17%) ⚠ mismatch sim=0.72

Capability score distribution (auxiliary)

  0-49     █████████████████████░░░░    5
  50-69    ░░░░░░░░░░░░░░░░░░░░░░░░░    0
  70-89    ████░░░░░░░░░░░░░░░░░░░░░    1
  90-100   ░░░░░░░░░░░░░░░░░░░░░░░░░    0

Host integrity ranking

Host	Runs	Average score	Family match	Most common V3 match
https://rhz4zlv.abc-tunnel.us/v1	2	0	0%	—
https://api.gnrt.dev/v1	2	41	0%	—
https://api.owgwen.com/v1	1	0	0%	—
https://api.cosmoshub.tech/v1	1	0	0%	anthropic/claude-opus-4.7 (1)

Recent anomalies

https://rhz4zlv.abc-tunnel.us/v1Aug 18, 2026, 4:36 AM

Self-claimed model: qd/ultimate
V2 predicted family: anthropic
V3 best match: —
Verdict: V3 low confidence · Abstain

https://rhz4zlv.abc-tunnel.us/v1Aug 18, 2026, 4:28 AM

Self-claimed model: qd/ultimate
V2 predicted family: anthropic
V3 best match: —
Verdict: V3 low confidence · Abstain

https://api.gnrt.dev/v1Aug 11, 2026, 9:27 AM

Self-claimed model: qd/ultimate
V2 predicted family: anthropic
V3 best match: —
Verdict: V3 low confidence · Abstain

https://api.owgwen.com/v1Aug 10, 2026, 2:43 AM

Self-claimed model: sr/qd/ultimate
V2 predicted family: anthropic
V3 best match: —
Verdict: V3 low confidence · Abstain

https://api.gnrt.dev/v1Aug 5, 2026, 5:22 PM

Self-claimed model: qd/ultimate
V2 predicted family: qwen
V3 best match: —
Verdict: Full swap

https://api.cosmoshub.tech/v1Jul 27, 2026, 2:43 PM

Self-claimed model: ultimate
V2 predicted family: anthropic
V3 best match: anthropic/claude-opus-4.7 (sim 0.72)
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 ultimate. Low family consistency, a different V3 match, or divergence from selfClaim is flagged as anomalous. BazaarLink Probe home.

Test an endpoint yourself →