Surge AI adds Mistral Large 4 to sudo L7 coding-agent benchmark

Surge AI has added Mistral AI's Mistral Large 4 to its sudo L7 benchmark leaderboard, which focuses on staff-level software engineering capabilities. The benchmark is designed to test whether coding agents can go beyond writing correct code and exercise the judgment of experienced engineers, such as navigating messy code bases, spotting hidden risks, and recognizing when technically correct code is still the wrong thing to ship.

According to the leaderboard, Mistral Large 4 scored 13.3%, ahead of Muse Spark 1.3's 12.2% and Grok 4.7's 10.6%. Surge AI referred to the model with the nickname “Le Chonk” and shared a link to the full leaderboard for developers to review the results.