Why the Sonnet Line Is Worth Watching
In Claude's product matrix, Opus is the heavy-hitting flagship, while Sonnet is the mid-tier workhorse—"capable enough, friendly on price, and faster"—which happens to be exactly the tier a huge amount of real workload actually runs on. So a Sonnet version update tends to hit closer to home for everyday users and pay-as-you-go developers than an update to the sky-priced flagship. The direction of 4.6's improvements is nothing surprising—coding, reasoning, tool calling, the core capabilities—but for a mid-tier line, the subtle shift in value (how much more capability, how much more speed, for the same money) is what matters.
An Upgrade to the Mid-Tier Line Is a Moment to "Re-Tier"
Every time the mid-tier model steps up a level, one question is worth re-asking: the work that used to require Opus—can Sonnet handle it now? Each such downward transfer of capability means a batch of tasks can move from the pricey tier to the cheaper one—and on high-volume automation pipelines, that price gap adds up fast within a single month. So the real way to use an update like Sonnet 4.6 isn't simply to "upgrade the default model," but to take your own task list and re-tier it once: which tasks drop down to the new Sonnet, which stay on Opus, and which can even go further down to Haiku. Teams that re-tier often have a billing curve shaped differently from everyone else's.
via: Hacker News