Bolt: Claude Sonnet 4.6 Integration
BoltBolt integrated Claude Sonnet 4.6 into its platform following Anthropic's February 17, 2026 release, making the new model available to all users via the agent selector. Sonnet 4.6 delivers near-flagship intelligence at mid-tier cost — users preferred it over Opus 4.5 roughly 59% of the time in head-to-head testing. The model's strongest upgrades are in agentic task execution: it plans multi-step features, parallelizes tool calls, and carries tasks through to completion with minimal hand-holding. As a launch incentive, Bolt offered Pro users a bonus of their full monthly token limit when using Sonnet 4.6 exclusively.
Sources & Mentions
5 external resources covering this update
Anthropic's Sonnet 4.6 matches flagship AI performance at one-fifth the cost, accelerating enterprise adoption
VentureBeat
Anthropic releases Sonnet 4.6
TechCrunch
Claude Sonnet 4.6
Hacker News
Claude Sonnet 4.6 model brings 'much-improved coding skills' and upgraded free tier
9to5Mac
Claude Sonnet 4.6 is now generally available in GitHub Copilot
GitHub
Claude Sonnet 4.6 Arrives in Bolt
Bolt added Claude Sonnet 4.6 to its platform on February 17, 2026 — the same day Anthropic released the model publicly. Users can switch to it from the agent selector inside Bolt, positioning Sonnet 4.6 alongside the previously integrated Opus 4.6 as the second major model addition in Bolt's February 7–20 release window.
What Sonnet 4.6 Brings to Bolt Builds
Sonnet 4.6 is, in Anthropic's framing, a "full upgrade" across coding, computer use, long-context reasoning, and agentic planning — not an incremental patch. For Bolt users, the most immediately felt improvements are in how the model handles complex, multi-step build tasks. Sonnet 4.6 proactively breaks work into parallel streams and pushes through to completion without requiring repeated prompts to stay on track. Anthropic's own testing found that developers preferred Sonnet 4.6 over Opus 4.5 — the previous flagship — approximately 59% of the time in Claude Code sessions, with users citing fewer hallucinations, less over-engineering of solutions, and more consistent follow-through.
The model also ships with a 1 million token context window in beta, matching the capacity of Opus 4.6. For Bolt projects involving large or rapidly growing codebases, this means the agent can maintain full awareness of project state without losing context mid-session.
Pricing and Performance Position
Sonnet 4.6 is priced at the same rate as Sonnet 4.5 — $3 per million input tokens and $15 per million output tokens via the API — while delivering performance that previously required reaching for Opus-class models. VentureBeat noted it matches flagship AI performance at roughly one-fifth the cost, which translates directly to more efficient token usage for Bolt users on metered plans.
Launch Promotion
To mark the release, Bolt ran a 48-hour promotion giving Pro plan users a bonus equal to their full monthly token allocation when building exclusively with Sonnet 4.6. Free users also received a promotional token bump during the same window. The offer was announced directly on Bolt's X account.
Choosing Between Sonnet 4.6 and Opus 4.6
With both Sonnet 4.6 and Opus 4.6 now available in Bolt, developers have a meaningful choice to make based on task type and token budget. Bolt's model picker allows switching at any point during a session. Sonnet 4.6 is well-suited to the bulk of everyday build tasks — feature implementation, refactoring, debugging — where its speed and cost efficiency outweigh the incremental capability edge of Opus. Opus 4.6 remains the appropriate choice for the most demanding reasoning tasks, particularly those requiring deep architectural decisions or extensive parallel agent work across very large codebases.