List your company

Multi-turn Code Generation on MT-Bench Coding

9First Turn ScoreGPT-4
3.0724.6116.157.689Feb 22, 2024

Evaluation results

MethodLinks
GPT-42024.0298.18.6
OpenCodeInterpreterBackbone=DeepSeekCoder-33B2024.026.86.76.8
OpenCodeInterpreterBackbone=DeepSeekCoder-6.7B2024.026.75.86.3
DeepSeekCoder-33B-Instruct2024.026.74.25.5
GPT-3.5-Turbo2024.026.67.26.9
Claude-Instant-V12024.026.576.8
OpenCodeInterpreterBackbone=CodeLlama-70B2024.025.73.74.7
DeepSeekCoder-6.7B-Instruct2024.025.64.55.1
OpenCodeInterpreterBackbone=CodeLlama-34B2024.025.64.14.9
OpenCodeInterpreterBackbone=CodeLlama-13B2024.025.53.74.6
CodeLlama-34B-Instruct2024.024.53.44
Vicuna-33B-V1.32024.023.82.93.4
Vicuna-13B-V1.32024.023.82.73.3
CodeLlama-7B-Instruct2024.023.32.52.9
CodeLlama-13B-Instruct2024.023.32.32.8