Model
Mistral Codestral 2 Achieves Top Coding Benchmarks, Challenges GitHub Copilot Economics
Mistral AI releases Codestral 2, achieving state-of-the-art results on HumanEval and SWE-bench. Analysis of how this open model changes the economics of AI coding assistants and threatens closed-source incumbents.
2025年4月13日来源:VentureBeat AI
mistralcodestralcoding-aibenchmarkopen-source