Breaking

Anthropic Releases Claude Sonnet 5.5 as Faster Mid-Tier Model Near Opus 5.5 on Work Benchmarks

September 28, 2026 at 17:46 EDT

Foundation Models
AI Agents
Software Dev & Coding

Anthropic released Claude Sonnet 5.5 on September 28, positioning the model as a faster, lower-cost complement to Claude Opus 5.5 for coding, document work, interface design and other well-scoped professional tasks. The company says the new Sonnet runs more than 30% faster than Claude Sonnet 5 and can cost up to 30% less per task because it uses fewer tokens and tool calls, while keeping the same API price as its predecessor.

September 28, 2026 · Anthropic

Sonnet 5.5 Moves Into Opus Territory

Claude Sonnet 5.5 is positioned as a faster, cheaper workhorse for coding and everyday professional tasks, with benchmark scores close to Claude Opus 5.5 in several areas.

Terminal-Bench 4.0: the coding jump

70.6%

Sonnet 5.5

10.3%

Sonnet 5

66.4%

Opus 5.5

The new Sonnet score is shown to scale against its predecessor and Opus 5.5.

30%+

faster output than Sonnet 5

per 1M input tokens

1844

GDPval-AA score

Where Anthropic draws the line

Sonnet 5.5

Well-scoped coding, bug fixes, documents, slides, spreadsheets and interface refinement.

Opus 5.5

Complex, ambiguous work that needs sustained judgment over longer assignments.

Early positives

Developers and enterprise testers reported faster iteration, fewer tool calls, lower token use and stronger routine coding performance.

Limits remain

Anthropic still positions Opus 5.5 as the better model for open-ended work, and some higher-risk cybersecurity tasks may fall back to Sonnet 5.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.