# Bolt's agent now runs on Claude Sonnet 5 By Steven Van · 2026-06-30 The upgrade brings stronger coding and edge-case handling to Bolt's full-stack app generation, with rollout beginning shortly. Bolt's AI agent just got a serious upgrade — **Claude Sonnet 5** is now part of the engine powering your builds, with rollout to users beginning shortly. [Source post](https://x.com/i/status/2072020128930156599) ## Sonnet 5 Joins Bolt's Agent Stack Claude Sonnet 5 is Anthropic's most agentic Sonnet model to date — it makes plans, uses tools like browsers and terminals, and can run tasks autonomously at a level that until recently required larger, more expensive models. Bringing that capability into Bolt means the agent powering your app generation gets a significant step up in reasoning and task completion. ## What Changes for Your Builds Sonnet 5's strength is in coding and catching edge cases — it handles more complex prompts with greater reliability and is particularly sharp at flagging risky or ambiguous requests before they become problems in your app. For Bolt users building full-stack apps with databases, auth, and deployment, that translates to fewer back-and-forths and more accurate first-pass results. ## Why It Matters For creators building production-ready apps in Bolt, this upgrade means the agent at the core of your workflow is now operating at a genuinely new capability level — not just faster, but more capable of seeing a complex build through from prompt to working product without losing the thread. ![Bolt]() Bolt AI coding platform that turns text prompts into full-stack websites and apps — with built-in databases, auth, hosting, and analytics. [View Bolt →]() Sources [the announcement on X](). [Read Bolt's agent now runs on Claude Sonnet 5 on Creators Toolbox]() --- Canonical source: https://creatorstoolbox.com/blog/bolt-new-sonnet-5-in-bolt