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

Code Generation on BigCodeBench-Completion Full

59.7pass@1

DeepSeek-Coder-V2-Instruct

24.3433.5242.751.88Dec 19, 2025Dec 24, 2025Dec 30, 2025Jan 5, 2026Jan 10, 2026Jan 16, 2026Jan 22, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.01
59.7
2025.12
59.7
2025.12
58.6
2025.12
58
2025.12
56.7
2025.12
55.4
54.8
2025.12
53.9
2025.12
53.6
53.3
53.2
2026.01
52.5
52.3
52.2
2025.12
52
2025.12
51.8
2026.01
51.7
51.1
2025.12
51.1
2026.01
50.9
2025.12
50.9
2026.01
49.6
2026.01
49
48.8
2025.12
48.8
47.6
2025.12
47.6
2025.12
45.8
45.5
45.4
45.1
2025.12
45.1
43.8
2025.12
43.8
2026.01
43.6
2026.01
41.6
2026.01
40.5
2025.12
35.6
2025.12
31.7
2026.01
25.7
2025.12
25.7