Model
Fraud Detection Net
Version v2.7
Low drift
Accuracy
91.6%
Latency
204ms
Cost / call
$0.002
7-day trend
+0.2
holding or improving
Drift timeline
Monthly accuracy for Fraud Detection Net over the last 12 months.
OctNovDecJanFebMarAprMayJunJulAugSep
| Month | Accuracy | Change | Drift |
|---|---|---|---|
| Sep 2026 | 91.6% | -0.6 | Low |
| Aug 2026 | 92.2% | +0.3 | Low |
| Jul 2026 | 91.9% | +0.6 | Low |
| Jun 2026 | 91.3% | +0.7 | Low |
| May 2026 | 90.6% | +0.2 | Low |
| Apr 2026 | 90.4% | -0.2 | Low |
| Mar 2026 | 90.6% | -0.5 | Low |
| Feb 2026 | 91.1% | -0.3 | Low |
| Jan 2026 | 91.4% | +0.2 | Low |
| Dec 2025 | 91.2% | +0.7 | Low |
| Nov 2025 | 90.5% | +0.7 | Low |
| Oct 2025 | 89.8% | — | Medium |
Accuracy
Evaluation suites
Regression — golden set98% pass
Nightly, 12k cases · ran 2026-09-13
Robustness — adversarial91% pass
Perturbation suite v3 · ran 2026-09-13
Fairness — group parity82% pass
18 of 100 cohorts flagged · ran 2026-09-13
Hallucination — grounded QA74% pass
RAG benchmark, 2k prompts · ran 2026-09-13
Champion vs challenger
No challenger is running against this model.
Governance
- Registered asset
- Fraud Detection Net
- Compliance score
- 67