Claude Sonnet 4.6 makes the cheap model question more interesting. Anthropic describes Sonnet 4.6 as a major upgrade for coding, computer use, long-context reasoning, agent planning, knowledge work, and design. It also has a 1M token context window in beta and is now the default for free and pro Claude users. The detail I care about is not only that it is better. It is that stronger capability is moving into the more commonly used model tier.
Developers love reaching for the strongest model available. I get it. I do it too when the task is hard enough. But in real teams, model choice is a tradeoff between cost, latency, context length, reliability, tool support, and task difficulty. You do not need the most expensive model for every test fix, small refactor, documentation update, or straightforward analysis task.
The mature version of this is model routing. Use the smaller or cheaper model where it is good enough, then escalate when the task genuinely needs deeper reasoning, longer planning, or more careful review. That is not just about saving money. It is about designing the workflow properly.
The interesting question is not “which model is best?” The better question is “which model is right for this job?” Claude Sonnet 4.6 makes that question more interesting because it raises the floor of what a more accessible model can do. That is good for developers, but it also means teams need better habits around choosing tools deliberately.