Breaking

Anthropic launches Claude Sonnet 5.5 as a faster, lower-cost counterpart to Opus 5.5

September 29, 2026 at 12:16 EDT

Foundation Models
AI Agents
Software Dev & Coding

Anthropic released Claude Sonnet 5.5 on September 28, 2026, positioning it as the faster, lower-cost counterpart to Claude Opus 5.5 for everyday coding, document, and business-agent workloads. The company says Claude Sonnet 5.5 is 30%+ faster than Claude Sonnet 5, costs up to 30% less per task, and raises Terminal-Bench 4.0 performance to 70.6% from 10.3%.

Sept. 28, 2026 - Anthropic

Sonnet takes the speed-and-cost role

Claude Sonnet 5.5 is pitched as the faster, lower-cost partner to Opus 5.5, with the biggest jump showing up in agentic coding.

70.6%

Terminal-Bench 4.0

30%+

faster output than Sonnet 5

$2/$10

input/output per MTok

token context window

Terminal-Bench 4.0: the headline jump

Each stacked block represents roughly 10 percentage points.

10.3%

Sonnet 5

66.4%

Opus 5.5

70.6%

Sonnet 5.5

Where it fits

Routine coding, document work, support flows, and high-volume review passes where speed and task cost matter.

Where it is weaker

Hard, open-ended judgment still favors Opus 5.5, and higher effort settings can spend many more tokens.

Adoption watchpoints

Effort levels
Lower settings save time and tokens; higher settings reason longer.

Migration
Thinking-off workflows move to between_tools.

Safeguards
Higher-risk cyber tasks can 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.