Accessibility Adjustments

Use these optional tools to adjust reading and display preferences. These tools cannot resolve every accessibility barrier. Please contact the website owner if you need assistance.

  • Text adjustments
  • Content scaling 100%
  • Font size 100%
  • Line height 100%
  • Letter spacing 100%
  • Colour adjustments
  • Orientation adjustments

Anthropic sets November 30 deadline for Claude Sonnet 4.5

Claude Sonnet 4.5 remains available until its November 30 retirement. Anthropic recommends Sonnet 5.5 and gives developers a migration checklist.

Listen to this article

Anthropic has set November 30 as the retirement date for Claude Sonnet 4.5 on the Claude API, giving developers a migration deadline for applications that still depend on the older model. The notice was issued on September 30 and recommends Claude Sonnet 5.5 as the replacement.

For teams running customer support, document processing or coding workflows, the immediate task is to find where the old model is still being called. A working integration today can become a failed request when the retirement date arrives.

What changes on November 30

The official deprecation notice lists Sonnet 4.5 as deprecated. It remains functional for now, but Anthropic no longer recommends it. Once a model is retired, requests to it fail.

The published schedule covers Anthropic operated platforms. Amazon Bedrock and Google Cloud maintain their own retirement schedules, so teams using those services need to check the corresponding provider notice rather than assume the same deadline applies everywhere.

Anthropic also points customers to the usage export in Claude Console, which breaks activity down by API key and model. That provides a starting point for identifying forgotten jobs and less visible dependencies.

The replacement needs testing

Sonnet 5.5 arrived on September 28. In its launch announcement, Anthropic positioned it as the faster, lower cost companion to Opus 5.5 for well defined work, including bug fixes and document creation. The company reported output speeds more than 30 percent above Sonnet 5 and costs up to 30 percent lower per task in its own testing.

Those comparisons are with Sonnet 5. They do not establish what a particular Sonnet 4.5 application will cost after migration. Teams should compare completed tasks, latency and failure rates on their own workload before committing to a replacement configuration.

Anthropic’s Sonnet 5.5 migration guide identifies changes beyond the model name. Adaptive thinking runs by default, response handling must account for different content block types, and older settings for thinking budgets, assistant prefills and forced tool use need review. A request format that worked with Sonnet 4.5 can therefore require application changes.

The guide includes a checklist organized by the model being replaced. Developers moving from Sonnet 4.5 should follow that section alongside the general requirements, rather than borrowing a migration example intended for Sonnet 5.

A shorter name still identifies a fixed version

There is also a naming change worth understanding. Anthropic’s versioning documentation says current model IDs without dates identify fixed releases. They are not automatic pointers to whichever model is newest. Earlier convenience aliases followed different rules.

That matters for change management. A team can select and test a specific replacement, record its configuration and monitor the result. It should still track that model’s separate support timeline instead of assuming the shorter name removes future migration work.

The practical next step is straightforward. Inventory active calls, test representative inputs and tool loops, and schedule the transition while there is time to investigate regressions. The November deadline is most useful when it becomes a planned maintenance task well before it becomes an outage.

Anthropic logo from LobeHub Icons, used under its MIT license.

Maya Chen
Maya Chen

Maya Chen is focused on covering AI models, research, and the evidence behind new capabilities. Maya follows model launches, benchmarks, open weights, and scientific uses of AI with one question in mind. What changed, and how would we know? The voice is curious and exacting, with a soft spot for elegant technical ideas and little patience for a leaderboard without context.

Leave a Reply

Your email address will not be published. Required fields are marked *

Gravatar profile