Bolt Adds Claude Sonnet 4.6 for Faster, Cheaper Agentic Coding.

Bolt announced the integration of Anthropic's Claude Sonnet 4.6 into its prompt-to-production platform. The new model gives users access to what Bolt describes as the most agentic Sonnet release yet, delivering stronger reasoning and proactive task execution at the same price point as Sonnet 4.5.

The timing is strategic. As AI-assisted development tools compete on both capability and cost, Bolt is positioning Sonnet 4.6 as the practical workhorse model for teams building full-stack applications daily, rather than reserving top-tier intelligence for occasional complex tasks.

According to Bolt, Sonnet 4.6 introduces three key improvements over previous Sonnet versions. First, it operates more autonomously, proactively executing tasks, planning multi-step features, and parallelizing work without requiring developers to guide every decision. Second, it handles complex front-end and back-end logic more reliably, maintaining coherence through longer development sessions. Third, it approaches Opus-level intelligence while maintaining Sonnet pricing, which matters for teams running hundreds of prompts per day.

The model's agentic capabilities align with Bolt's recent focus on autonomous development workflows. Following their Design System Agent launch, which generates production-ready prototypes from design tokens, Sonnet 4.6 adds another layer of proactive intelligence to the platform.

Bolt's pitch centers on economics. For teams accustomed to using Opus for everyday tasks, Sonnet 4.6 offers comparable reasoning at faster speeds and lower cost. The company didn't disclose specific pricing, but emphasized that Sonnet 4.6 maintains price parity with Sonnet 4.5 while delivering measurably better output quality.

This pricing strategy matters in the context of enterprise adoption. After partnering with Microsoft to bring Bolt to Azure and M365, the platform is clearly targeting organizations that need predictable costs at scale. Giving teams an Opus alternative that doesn't blow through token budgets makes AI-assisted development more sustainable for daily use.

The 48-hour promotional offer reflects this volume focus: Pro users receive their entire monthly limit as a bonus, while free users get double limits when using Sonnet 4.6 specifically. It's a clear push to drive adoption and gather usage data on the new model.

Bolt competes directly with Cursor and Windsurf in the AI coding space, but takes a different architectural approach. While Cursor and Windsurf integrate into existing IDEs, Bolt operates as a standalone web-based environment where users prompt, run, edit, and deploy full applications without leaving the browser.

The Sonnet 4.6 addition strengthens Bolt's position as a model-agnostic platform. Users can now choose between multiple Claude models depending on task complexity and budget, similar to how v0 and Lovable offer model selection for different use cases.

For developers already using Bolt's Connectors feature to pull live data from Notion, GitHub, and Linear, Sonnet 4.6 provides better reasoning over that external context, making it easier to build features that integrate with existing workflows.

Claude Sonnet 4.6 is available now at bolt.new for all Pro and Free users. The promotional bonus limits run for 48 hours from the February 17 announcement. Teams on Bolt's enterprise plan receive the same model access with additional security scans and admin controls shipped last year.