Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Code Generation on MBPP 3-shot (Accuracy, TPS, Steps, Latency)

59.6Accuracy

KLASS

24.44833.57442.751.826May 9, 2026May 12, 2026May 15, 2026May 18, 2026May 21, 2026May 24, 2026May 28, 2026
Updated 5d ago

Evaluation Results

MethodLinks
2026.05
59.69.2549.1
2026.05
56.438.12-0.88
2026.05
56.29.44-2.52
2026.05
5617.71-4.74
2026.05
5620.15-5.39
2026.05
5515121
2026.05
54.817.82718.1
2026.05
54.83.74-1
2026.05
54.6103410.1
2026.05
54.45.35-1.43
2026.05
54.229.06-3.17
2026.05
5420.64-5.52
2026.05
53.7333.2-3.62
2026.05
53.225.35-2.76
2026.05
5325.21825
2026.05
52.89.18-1
2026.05
49.217.02-1.85
2026.05
40.215.2677.6
2026.05
38.816.86-8.47
2026.05
38.61.99-1
2026.05
38.63.89-1.95
2026.05
3811.78-5.92
2026.05
37.416.27-8.18
2026.05
3715121
2026.05
3721.62220.3
2026.05
36.211.94511.3
2026.05
328.08-1.48
2026.05
31.9615.62-2.87
2026.05
31.820.01-3.67
2026.05
31.65.45-1
2026.05
30.626.53-4.87
2026.05
30.226.66-4.89
2026.05
25.81.45-0.73