In a recent experiment, three AI agents—Opus 4.7, GPT-5, and DeepSeek V4-Pro—were tasked with creating a TSS-ECDSA command-line interface using Rust. Each agent approached the challenge differently, showcasing their unique capabilities. Opus 4.7 completed the task in a remarkable 26 minutes, demonstrating speed and efficiency. In contrast, GPT-5 produced the most organized and readable code, emphasizing quality over speed. Meanwhile, DeepSeek V4-Pro appeared to overlook direct feedback during the development process, which may have impacted its performance. This comparison highlights the varying strengths of each AI, from rapid execution to code clarity, providing insights into their potential applications in software development.
Opus 4.7 vs GPT-5 vs DeepSeek V4-Pro: Three Agents Develop TSS-CLI in Rust
A comparison of three AI agents tackling the same Rust project reveals distinct strengths and weaknesses.
