Financial Compliance Classification on FinGuard-Bench response-level
85.43F1 ScoreFinGuard
Evaluation Results
| Method | Links | |
|---|---|---|
| FinGuardself-play=true2026.05 | 85.43 | |
| FinGuardself-play=false2026.05 | 84.71 | |
| Qwen3Guard-Gen-8Bmode=strict2026.05 | 73.49 | |
| GPT-5.12026.05 | 64.09 | |
| Qwen3.5-397B-A17B2026.05 | 64.08 | |
| Gemini-3.1-Flash2026.05 | 52.77 | |
| PolyGuard-Qwen-7B2026.05 | 46.08 | |
| Qwen3.5-9B2026.05 | 33.67 | |
| NemoGuard-8B2026.05 | 29.18 | |
| LlamaGuard3-8B2026.05 | 28.87 | |
| Qwen3-8B2026.05 | 22.91 | |
| Qwen3Guard-Gen-8Bmode=loose2026.05 | 19.42 |