PandasAI (YC W24) reposted this
🔥 Anthropic just released Claude 3.5 Sonnet: it's mindblowing 🔥 - 49% on SWE-bench Verified (up from 33.4%) - highest among all publicly available models - 69.2% on TAU-bench retail (up from 62.6%) - 46% on TAU-bench airline (up from 36%) - Outperforms specialized coding systems and even OpenAI's latest models This isn't just about numbers - I've been testing it extensively with PandasAI (YC W24), and the improvement in code quality and reasoning is remarkable. Early feedback from companies like GitLab shows up to 10% stronger reasoning across DevSecOps use cases, with no added latency. What makes this particularly exciting: - Industry-leading software engineering capabilities - Significantly improved reasoning for multi-step coding tasks - Same speed and pricing as the previous version - Exceptional performance in data analysis and pandas operations Have you tried the new Claude 3.5 Sonnet yet? I'd love to hear about your experience! Learn more: https://lnkd.in/dBtXZ2Tk