Code Infilling on HumanEval single-line
92.6Pass@1Qwen2.5-Coder-7B
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Qwen2.5-Coder-7BModel Type=Open-Weights AR2026.02 | 92.6 | — | — | |
| DREAMONModel Type=Open-Weights Diffusion, Backbone=DiffuCoder-7B2026.02 | 92.2 | — | — | |
| DREAMONModel Type=Open-Weights Diffusion, Backbone=DreamCoder-7B2026.02 | 92.1 | — | — | |
| Seed-Coder-8BModel Type=Open-Weights AR2026.02 | 89.7 | — | — | |
| DREAMONModel Type=Open-Weights Diffusion, Backbone=Dream-7B2026.02 | 88.6 | — | — | |
| Deepseek-Coder-6.7BModel Type=Open-Weights AR2026.02 | 73 | — | — | |
| DreamCoder-7BModel Type=Open-Weights Diffusion2026.02 | 55.5 | — | — | |
| DiffuCoder-7BModel Type=Open-Weights Diffusion2026.02 | 53.7 | — | — | |
| LLaDA-8BModel Type=Open-Weights Diffusion2026.02 | 48.3 | — | — | |
| Dream-7BModel Type=Open-Weights Diffusion2026.02 | 48.2 | — | — | |
| Code LlamaSize=7B, LCFT=false, Decoding Strategy=greedy decoding2023.08 | — | 77 | 83.3 | |
| Code LlamaSize=7B, LCFT=true, Decoding Strategy=greedy decoding2023.08 | — | 74.1 | 83.3 | |
| Code LlamaSize=13B, LCFT=false, Decoding Strategy=greedy decoding2023.08 | — | 80.7 | 85.9 | |
| Code LlamaSize=13B, LCFT=true, Decoding Strategy=greedy decoding2023.08 | — | 75.9 | 85.6 | |
| code-davinci-002Size=175B, Temperature=0.1, Decoding Strategy=nucleus sampling (p=0.95)2023.08 | — | 91.6 | — | |
| InCoderSize=6B, Temperature=0.2, Decoding Strategy=sampling2023.08 | — | 69 | — | |
| OpenAI FIM90Size=7B, Temperature=0.1, Decoding Strategy=nucleus sampling (p=0.95)2023.08 | — | 75.1 | — |