Claude Sonnet 5.5: Anthropic Launches a Sonnet 30% Faster at the Same Price as Sonnet 5

By Carlos Montiel | Enterprise AI Specialist
Leer en español →
Published: 2026-10-03 | By: Carlos Montiel | Reading time: ~4 minutes

Five days after Opus 5.5, Anthropic completed the second step of its 5.5 family: Claude Sonnet 5.5, the mid-tier model most companies run in production, arrived on September 29 with more speed and lower token consumption, without raising its list price.

What launched and where it's available

According to coverage from several outlets, Claude Sonnet 5.5 shipped on September 29, 2026 and is available from day one in Claude.ai (web, iOS, and Android) and through the API via Anthropic, AWS, Google Cloud, and Microsoft Foundry. Anthropic positions it as an efficiency upgrade rather than a radical leap in capability, aimed at everyday professional work: coding, writing, and office documents.

It is the second model in the 5.5 line after Claude Opus 5.5 (September 22). In that announcement the company said Sonnet 5.5 and Haiku 5.5 would arrive "in the coming weeks"; Sonnet has now delivered and, according to the press, Haiku 5.5 is still pending with no firm date.

The numbers we can actually cite

Launch date ................. Sep 29, 2026 Speed ....................... over 30% faster than Sonnet 5 Cost per task ............... up to 30% lower (fewer tokens per task) Reported list price ......... $2 / $10 per million tokens (input / output) Cache read .................. $0.20 per million tokens Reference: Opus 5.5 ......... $4 / $20 per million tokens

Two independent outlets we reviewed report the $2/$10 price and note it matches Sonnet 5; a third published $3/$15. Because the sources don't fully agree, confirm the current price on Anthropic's official page before budgeting. The "up to 30% per task" improvement doesn't come from a rate cut but from the model needing fewer tokens to finish the same job, a saving that depends on your actual workload.

About the benchmarks

The press reports that the model beats Opus 5.5 at agentic coding and matches Opus 5 on cybersecurity capabilities. We aren't citing benchmark percentages because the figures in circulation are inconsistent across sources and are the company's own measurements. To decide, the sensible move is to test it against your own use cases.

What it means for your company: Sonnet is the workhorse model for support agents, coding assistants, and document automation. With Opus 5.5 at double the list price, the practical question is which workflows truly need the big model. Recommendation: run a batch of 50 to 100 real tasks from your operation on both Sonnet 5.5 and Opus 5.5, compare quality and total cost per task (not just price per token), and route by default to the cheaper model that clears your quality bar. If you were already on Sonnet 5, migrating is a model-identifier change: test it in staging before switching production.
Carlos Montiel
Enterprise AI Solutions Architect
Specialist in LLMs, Agents, and Orchestration
guatemalia.com/#contacto · info@guatemalia.com

Need to implement AI in your company?

Carlos Montiel is an enterprise AI solutions architect. He implements LLMs, Agents, RAG, and orchestrators for companies across Guatemala and Latin America. Reach out for a consultation.

Contact Carlos Montiel

info@guatemalia.com