Claude is better for coding. ChatGPT is better for everything around the code. Claude Opus 5 leads SWE-bench Verified at 97.0% on the independent vals.ai leaderboard, against 88.6% for Opus 4.8, the model it replaced. GPT answers on agentic work, where GPT-5.6 Sol tops Artificial Analysis's Coding Agent Index at 80.2 Apr 2026