GPT-5.6 Luna vs. GPT-6 Astra: Is a $1.20 Model Good Enough for Code Review?

GPT-5.6 Luna, a significantly cheaper model at $1.20 per million output tokens compared to GPT-6 Astra's $50, found 75% as many verified bugs (69 vs. 92) across 50 code review pull requests while costing only 3.6% of the price. However, Luna had notably lower precision, with 24 of its 93 findings failing verification compared to only 4 of Astra's 96, and it detected fewer security bugs (9 vs. 19), leading the researchers to recommend Luna only for everyday correctness issues and not for security-sensitive code like authentication or permissions.