• •Developers on Hacker News report Claude 3.5 Sonnet consistently outperforms other models in nuance and adherence to complex coding styles.
• •ChatGPT o1 is noted for its reasoning capabilities, proving more effective at identifying deep-seated logic errors than its predecessors.
• •Users report inconsistent performance across both models when handling massive repositories exceeding typical context window limits.
• •The effectiveness of AI-led reviews remains subjective, as many developers continue to prefer human oversight for security-critical pull requests.

Software engineers are increasingly comparing Anthropic’s Claude 3.5 Sonnet and OpenAI’s ChatGPT o1 for their ability to conduct automated code reviews. While Claude is frequently cited by the Hacker News community for its stylistic accuracy, ChatGPT o1 is gaining traction for its complex logical reasoning. Both tools struggle with high-latency issues when processing large-scale codebases, leaving many teams to rely on a hybrid approach. Whether these tools will fully automate quality assurance depends on how well they integrate into existing CI/CD pipelines without introducing hallucinations.