Is anyone else doing just fine with more basic models in Claude Code?

A developer working with web apps and DevOps reports satisfaction with Claude's Sonnet and Haiku models, contrary to discussions about the latest Opus/Fable models being subpar. They find these basic models sufficient for their work, providing good output and test coverage, and have not encountered weekly limits. This prompts a question about whether others are also succeeding with less advanced technology.

This report offers a contrasting perspective, highlighting satisfaction with Claude's Sonnet and Haiku models, unlike the common discourse criticizing the latest Opus/Fable versions.