Fordel StudiosClaude Sonnet 4.6 released as Anthropic's new default model
What Happened
Anthropic released Claude Sonnet 4.6 on February 17, 2026 as the new default model for Claude Code, replacing Sonnet 4.5. In internal coding tests, engineers preferred Sonnet 4.6 outputs 70% of the time. The upgrade is automatic — no configuration change required for existing users.
Our Take
Okay, genuinely — free upgrades don't happen often in this industry. Usually 'new model' means 'new bill.' So when Anthropic just quietly makes Sonnet 4.6 the default and every Claude Code user gets it without touching a config, that's worth noting.
The 70% preference rate in coding tests is the number that matters here. Not a benchmark (those are cooked). Not a blog post. Actual engineers running actual tasks and picking the new one 7 out of 10 times. That's a real signal.
For a small shop like ours, Claude Code is basically a team member at this point. Better reasoning on multi-file refactors, fewer hallucinated APIs — that compounds fast across a sprint.
(The cynical read: this is also Anthropic quietly retiring Sonnet 4.5 without making a big deal of it. Which is fine. Just notice it.)
Look, I'm not going to pretend this is some paradigm-shattering — wait, no. It's just better. That's the whole story. And better-for-free is the best kind of better.
What To Do
Run your current Claude Code session against a task that previously needed multiple clarifications — Sonnet 4.6 is already active as the default, so you're testing it right now without doing anything.
Builder's Brief
What Skeptics Say
A 70% preference rate from internal Anthropic engineers is a self-reported benchmark with no external replication; real-world gains for non-coding workflows may be marginal and the model churn creates prompt-regression risk for teams with tuned system prompts.
Cited By
React
Get the weekly AI digest
The stories that matter, with a builder's perspective. Every Thursday.