Skip to content
NewsModel release

Anthropic ships Claude Sonnet 5.5, nearly matching Opus 5.5 at half its API price

· by Pondero Newsdesk

The short version

Anthropic's new mid-tier model scores within 2 points of Opus 5.5 on real-world work benchmarks while costing half as much per token, and GitHub Copilot added it the same day.

Anthropic ships Claude Sonnet 5.5, nearly matching Opus 5.5 at half its API price

Anthropic released Claude Sonnet 5.5 on September 28, 2026, and on several of the company's own benchmarks the mid-tier model now scores within a couple of points of Opus 5.5, the flagship it released the week before, at half the per-token price, per Anthropic's release notes.

What

Sonnet 5.5 is the second model in the Claude 5.5 family. Anthropic says it generates output more than 30 percent faster than Sonnet 5 and needs far fewer tokens for the same work, cutting cost per task by up to 30 percent, while API pricing holds at $2 per million input tokens and $10 per million output tokens, with $0.20 per million tokens for cache reads and $2.50 for cache writes, per Anthropic. On Anthropic's internal Terminal-Bench 4.0 coding evaluation, the score jumped from 10.3 percent for Sonnet 5 to 70.6 percent for Sonnet 5.5, ahead of Opus 5.5's reported 66.4 percent on the same settings. The model is the first Sonnet release to ship with the cyber safeguards and fallback protections Anthropic previously reserved for its most capable models, and the first with safety classifiers Anthropic says block reasoning-extraction attempts used in distillation attacks. It is available now on the Claude Platform, AWS, Google Cloud and Microsoft Azure with zero data retention, per Anthropic. GitHub added the model to Copilot the same day, generally available to Pro, Pro+, Max, Business and Enterprise users across VS Code, Visual Studio, the Copilot CLI, JetBrains, Xcode and github.com, billed at provider list pricing under usage-based billing, per GitHub's changelog.

Sonnet 5.5 now covers work that used to require paying Opus prices

On GDPval-AA, a test of real-world tasks across 44 occupations, Sonnet 5.5 scored 1844 against Opus 5.5's 1846 and Sonnet 5's 1449, and on AA-Briefcase it scored 1811 against Opus 5.5's 1822, per Anthropic's benchmark page. Opus 5.5 costs $4 per million input tokens and $20 per million output tokens, twice Sonnet 5.5's rate, VentureBeat reported. That narrows the practical case for defaulting to Opus on everyday coding, document and spreadsheet work, the tasks Anthropic now explicitly steers toward Sonnet. Early enterprise testing backs the efficiency claim: Box's VP of AI Products said the model ran 2.4 times faster and used 12 percent fewer tokens than its predecessor, Zendesk said support tickets processed 20 percent faster, and Slack reported roughly 14 percent fewer output tokens on its Slackbot evaluations without any prompt changes, Anthropic told VentureBeat. For teams routing between Claude models by cost, Sonnet 5.5 is now the model to try first before reaching for Opus.

Anthropic is not claiming Sonnet 5.5 closes the gap everywhere. The company says Opus 5.5 remains clearly stronger on complex, open-ended work requiring sustained judgment, and Sonnet 5.5 still trails on CursorBench 4.0, 55.5 percent against Opus 5.5's 57.8 percent, per Anthropic.

What to watch next

Anthropic says Claude Haiku 5.5 is coming "in the coming weeks" to round out the 5.5 family for high-volume, cost-sensitive applications, per Anthropic's release notes. GitHub's rollout to Copilot users is gradual, so seat access may lag the announcement by days.

Sources