Verification on VerifyBench Hard 1.0 (test)
91.9Mean@3 AccuracyCoSineVerifier-Tool-4B
Evaluation Results
| Method | Links | |
|---|---|---|
| CoSineVerifier-Tool-4BVerifier Type=CoT Verifier, Model Access Category=Open-source2025.12 | 91.9 | |
| CoSineVerifier-32BVerifier Type=Labeling Verifier2025.12 | 90 | |
| CompassVerifier-32BVerifier Type=Labeling Verifier2025.12 | 88.9 | |
| o3Verifier Type=CoT Verifier, Model Access Category=Closed-source2025.12 | 88.7 | |
| Qwen3-235B-A22B-2507Verifier Type=CoT Verifier, Model Access Category=Open-source2025.12 | 87.7 | |
| CompassVerifier-32B (CoT)Verifier Type=CoT Verifier, Model Access Category=Open-source2025.12 | 86.5 | |
| Gemini2.5-FlashVerifier Type=CoT Verifier, Model Access Category=Closed-source2025.12 | 86 | |
| CoSineVerifier-4BVerifier Type=Labeling Verifier2025.12 | 85.4 | |
| Qwen3-32BVerifier Type=CoT Verifier, Model Access Category=Open-source2025.12 | 85.2 | |
| CompassVerifier-7BVerifier Type=Labeling Verifier2025.12 | 85.2 | |
| GPT-oss-20BVerifier Type=CoT Verifier, Model Access Category=Open-source2025.12 | 84.7 | |
| GPT-4oVerifier Type=CoT Verifier, Model Access Category=Closed-source2025.12 | 84.6 | |
| Qwen3-4B-Instruct-2507Verifier Type=CoT Verifier, Model Access Category=Open-source2025.12 | 84.1 | |
| Qwen3-8BVerifier Type=CoT Verifier, Model Access Category=Open-source2025.12 | 83.6 | |
| XVerify-8B-IVerifier Type=Labeling Verifier2025.12 | 83.3 | |
| CompassVerifier-7B (CoT)Verifier Type=CoT Verifier, Model Access Category=Open-source2025.12 | 82.6 | |
| Qwen3-4BVerifier Type=CoT Verifier, Model Access Category=Open-source2025.12 | 80.3 | |
| LLaMA3.3-70B-InstructVerifier Type=CoT Verifier, Model Access Category=Open-source2025.12 | 77.2 |