Claude Opus 5.5: Keeping safety ahead of capabilities

https://t2.gstatic.com/faviconV2?client=SOCIAL&type=FAVICON&fallback_opts=TYPE,SIZE,URL&url=https://anthropic.com&size=128

Anthropic today introduced Claude Opus 5.5, the first release in a new family of Claude 5.5 models. Claude Opus 5.5 performs at the level of Claude Fable 5.1 on most work, and costs around 40% less to run than Opus 5. Over the coming weeks, we’ll also be launching Claude Sonnet 5.5 and Haiku 5.5.

This is the first model release since the company called for pacing the frontier, which means keeping safety practices ahead of model capabilities.

What this means for our model development and release

The current generation of models relies on extensive alignment testing, pre-release evaluation by outside organizations such as METR and Frontier Design, and safeguards matched to each model’s capabilities in high-risk areas like cybersecurity and biology.

For Opus 5.5 specifically,

Anthropic MCP Anthropic

The model context protocol for AI connectivity

AI Gateway & Proxy Compare pricing & 5 other AI Gateway & Proxy →

Anthropic ran alignments tests that show Opus 5.5 is the strongest performing model to date. Noteworthy are improvements to cybersecurity behaviors such as biased reasonbing and attempting to escape a sandbox, among others. It is also the first Opus model with similar cybersecurity and biology safeguards as Fable 5.1. According to Anthropic’s announcement, “When the safeguards intervene, requests fall back to another model transparently. For cyber, users will be able to identify and fix bugs in their code, but most cybersecurity tasks will be re-routed to Opus 4.8. Requests flagged by biology classifiers will be re-routed to Opus 5.”

As for pricing, Opus 5.5 requires less compute to serve than Opus 5, the company announced, and will cost 40% less than Opus 5 on typical workloads. Input and output tokens are $4 and $20 per million, 20% less than Opus 5., Anthropic said.

Mario Rodriguez, chief product officer at GitHub, said “Developers want agents that can take on real software work and finish it. In our testing across GitHub Copilot CLI and VS Code, Claude Opus 5.5 used among the fewest tokens and steps we measured. In VS Code, it solved more terminal tasks than Opus 5 in less than half the steps. More than making individual tasks more efficient, it’s making developers’ bigger projects more achievable.”

Claude Opus 5.5 is now available on all platforms, including Amazon Web Services, Google Cloud, and Microsoft Azure. On the Claude Platform, developers can get started with Claude Opus 5.5.

To learn more, read Anthropic’s blog announcement.

The post Claude Opus 5.5: Keeping safety ahead of capabilities appeared first on SD Times.

Tech Developers