Dario Amodei, CEO of Anthropic, attends the AI Impact Summit in New Delhi, India on Thursday, February 19, 2026.
Ruhani Kaur | Bloomberg | Getty Images
Anthropic announced Thursday its new artificial intelligence model Claude Opus 4.7. The company says it’s an improvement over past models, but has a “broader range of features” than its latest product, Claude Mythos Preview.
Claude Opus 4.7 excels at software engineering, following instructions and completing real-world tasks, and is the most powerful model available to the public, Anthropic said. However, the model’s cyber capabilities are not as advanced as the Claude Mythos Preview that Anthropic rolled out to a select group of companies earlier this month as part of a new cybersecurity initiative called Project Glasswing.
“We are releasing Opus 4.7 with protections that automatically detect and block requests that indicate prohibited or high-risk cybersecurity usage,” Anthropic said in a release. “What we learn from the actual deployment of these safety measures will help us work toward our ultimate goal of broadly releasing the Mythos class of models.”
Founded in 2021, Anthropic has spent years carefully building a reputation as a company dedicated to safer and more responsible AI deployments than rivals such as OpenAI.
The launch of Project Glasswing led to a number of high-profile meetings between members of the Trump administration, technology CEOs, and bank CEOs about the security risks of powerful AI models.
Anthropic has no plans to make the Claude Mythos preview publicly available, but the company says its goal is to eventually learn how to deploy Mythos class models at scale.
Thursday’s release of Claude Opus 4.7 comes after Anthropic released Claude Opus 4.6 in February. Anthropic said in a release that the new model outperforms Claude Opus 4.6 in many use cases, including industry benchmarks such as agent coding, interdisciplinary reasoning, use of extension tools, and use of agent computers.
Anthropic said it experimented with efforts to “differentially degrade” Claude Opus 4.7’s cyber capabilities during the exercise.
The company encouraged security professionals interested in using the model for “legitimate cybersecurity purposes” to apply through a formal validation program.
Claude Opus 4.7 is available through all of Anthropic’s Claude products, application programming interfaces, and cloud providers. microsoft, google and Amazon. Anthropic says the new model will cost the same as the Claude Opus 4.6.
WATCH: Anthropic co-founder says company is expanding access to Mythos: Semafor event
