Breaking

Anthropic Releases Claude Sonnet 5.5 With Faster Output and Lower Task Costs

September 28, 2026 at 14:08 EDT

Foundation Models
AI Agents
Software Dev & Coding

Anthropic released Claude Sonnet 5.5 on September 28, updating its mid-tier Claude model with faster output, lower task cost and a stronger focus on everyday agent work. The company says the new model runs more than 30% faster than Claude Sonnet 5 and costs up to 30% less for most work, even though the per-token API price stays the same.

September 28, 2026 · Anthropic

Claude Sonnet 5.5 speeds up the middle tier

Anthropic's new Sonnet model is built for well-scoped coding, bug fixes, documents and fast agent iteration, with lower task cost rather than lower per-token prices.

30%+

faster output than Sonnet 5

30%

less cost per task for most work

70.6%

Terminal-Bench 4.0 score

$2/$10

input and output price per MTok

Terminal-Bench 4.0 jump

10.3%

Sonnet 5

70.6%

Sonnet 5.5

Takeaway: The biggest visible benchmark change is in agentic command-line coding.

Where the gains come from

Scoped task

→

Batched tool calls

→

Fewer steps

→

Lower task cost

The API price stays at $2 input and $10 output per MTok; the savings come from faster generation and fewer tokens needed to finish work.

Best fit

Routine implementation, bug fixing, UI polish, support workflows, documents, slides and spreadsheets.

Watchouts

Breaking API changes, visible cyber fallbacks and possible false flags in narrow biology areas require migration testing.

Continue reading

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