Efficient model (128k tokens) for general coding, Q&A, and web-enabled answers with fast response.