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

Code Generation on HumanEval (Throughput, Speedup, and Avg. τ)

1.94Average Tau (τ)

MEDUSA

1.58164.00086.428.8392Dec 17, 2025Dec 25, 2025Jan 2, 2026Jan 11, 2026Jan 19, 2026Jan 27, 2026Feb 5, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2025.12
1.94-1.34
2025.12
1.97-1.32
2025.12
2.53-2.3
2025.12
2.57-2.14
2025.12
2.58-2.39
2025.12
2.61-1.91
2025.12
2.63-2.01
2026.02
3.01-1.74
2026.02
3.05-1.84
2026.02
3.05-1.75
2026.02
3.17-1.89
2025.12
3.21-2.41
2025.12
3.23-2.18
2025.12
3.25-2.41
2025.12
3.26-2.47
2025.12
3.34-2.65
2026.02
3.39-1.94
2026.02
3.47-2.12
2026.02
3.54-2.05
2026.02
3.65-2.17
2025.12
3.72-2.62
2025.12
3.87-2.65
2025.12
3.9-2.52
2025.12
4.21-2.79
2025.12
4.51-2.59
2026.02
4.5722.53-
2025.12
4.66-3.18
2025.12
4.68-2.97
2025.12
5-3.3
2025.12
5-3.5
2026.02
5.1321.17-
2026.02
5.3921.2-
2025.12
5.48-4.32
2026.02
5.52-4.32
2025.12
5.72-3.59
2026.02
6.04-4.74
2026.02
6.59554.2
2026.02
6.5-5.14
2025.12
6.58-5.54
2026.02
6.631,2474
2026.02
6.64-5.21
2026.02
7.4624.2-
2026.02
8.098023.5
2026.02
8.3649.052.71
2026.02
10.951.072.82
2026.02
-312-
2026.02
-229-
2026.02
-229-
2026.02
-18.11