Code Generation on MBPP (accuracy, throughput, speedup)
60Accuracy
Latent Thinking Optimization
Updated 2d ago
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
2025.09 | 60 | - | - | |
2026.01 | 58.8 | 4.7 | 1 | |
2025.09 | 57 | - | - | |
2026.01 | 56.2 | 4.7 | 1 | |
2025.09 | 56 | - | - | |
2026.01 | 55.4 | 31.1 | 6.6 | |
2026.01 | 54.4 | 29.7 | 6.3 | |
2026.01 | 53.8 | 15.9 | 3.4 | |
2026.01 | 53.2 | 25.3 | 5.4 | |
2026.01 | 53.2 | 6.4 | 1.4 | |
2026.01 | 51 | 10.1 | 2.2 | |
2026.01 | 50.2 | 24.9 | 5.3 | |
2026.01 | 48.6 | 10 | 2.1 | |
2025.07 | 39.2 | - | - | |
2025.07 | 36.8 | - | - | |
2025.07 | 36.8 | - | - | |
2025.07 | 36.6 | - | - | |
2025.07 | 36.4 | - | - | |
2025.07 | 35.6 | - | - | |
2025.09 | 32.2 | - | - | |
2025.09 | 26.3 | - | - | |
2025.09 | 24.7 | - | - | |
2025.09 | 23.7 | - | - | |
2025.09 | 19.3 | - | - | |
2025.09 | 18.9 | - | - |