Claude Sonnet 5.5 highlights the importance of cost per completed task

What changed Anthropic introduced Claude Sonnet 5.5 on September 28, 2026. The company positions the model for well-defined coding tasks, bug fixes, and document work, alongside its higher-end Opus offering. Anthropic reports faster output and lower task costs than Sonnet 5 in its own testing. Those are vendor claims, not measurements performed for this article. The announcement makes a distinction that matters when comparing AI systems: a model's price per token is not the same as the price of completing a task. A cheaper answer can still be expensive if it requires several retries and substantial manual correction. Our perspective For a software team, a useful comparison includes whether the patch passes tests, how much review it needs, and whether it follows the project's conventions. Keep prompts and acceptance criteria consistent across models, and record failures as carefully as successful demonstrations. A practical rollout can start with low-impact work such as drafting tests or explaining unfamiliar code. Expand the scope when the results justify it. Preserve human review for changes that affect authentication, user data, or production infrastructure. There is no single best model for every workload. The stronger choice is the one that reliably finishes your particular tasks within the time and budget available. Announcement: September 28, 2026. Summary and analysis published November 3, 2026. Source https://www.anthropic.com/claude-sonnet-5-5