> Markdown version of [/events/world-congress-2025/sessions/723-evaluating-ai-models](https://www.wearedevelopers.com/events/world-congress-2025/sessions/723-evaluating-ai-models). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md). --- # Evaluating AI models for code comprehension - **Date:** Friday, Jul 11, 2025 - **Time:** 09:00–09:30 (30 min) - **Room:** Stage 11 - **Event:** World Congress 2025 - **Tags:** ai, devops, llms, nlp, performance, productivity ## Recording [Watch recording](https://www.wearedevelopers.com/videos/1462-evaluating-ai-models-for-code-comprehension) ## Description With so many AI models out there, which one is actually the best at understanding code? As we've built Graphite Reviewer, our high-signal AI-powered code review companion, we've run thousands of tests, gotten tons of feedback from users, and have extensively examined how these various models perform in real world scenarios. In this talk we'll explore the nuances between these models, how they perform specifically in relation to code comprehension, and finally answer the question: which AI model should developers be using? This talk will include a deep dive into the various experiments we've run on AI code comprehension, comparing all of the top AI models today, and an exploration of the results and data we’ve collected. ## Speaker ### [Merrill Lutsky](https://www.wearedevelopers.com/@merrill-lutsky) Graphite CEO and Co-founder ## Related talks at this congress - [How we built an AI-powered code reviewer in 80 hours](https://www.wearedevelopers.com/events/world-congress-2025/sessions/917-how-we-built-an-ai) — Yan Cui - [AI-Powered Code Documentation: Simplify the Complex](https://www.wearedevelopers.com/events/world-congress-2025/sessions/719-ai-powered-code) — Patrick Schnell - [Leapter: The Reinvention of Software Development? A Future Built On AI Generated Code.](https://www.wearedevelopers.com/events/world-congress-2025/sessions/581-leapter-the) — Robert Werner - [Bringing AI Model Testing and Prompt Management to Your Codebase with GitHub Models](https://www.wearedevelopers.com/events/world-congress-2025/sessions/638-bringing-ai-model) — Kevin Lewis, Sandra Ahlgrimm