AgentCoder: Multi-Agent-based Code Generation with Iterative Testing and Optimisation

Open paper
Benchmark Model Rank Results
code-generation-on-mbppGPT-4 + AgentCoder#4Accuracy: 91.8
code-generation-on-mbppGPT-3.5 Turbo (ChatGPT) + AgentCoder#6Accuracy: 89.9

Rank counts only results with a code link.