Anthropic releases Claude Opus 4.7, pushing its coding and vision capabilities further
Anthropic has released Claude Opus 4.7, a new flagship model the company says is now generally available across its consumer products, API, and major cloud platforms.
The launch marks a notable step up from Opus 4.6, particularly in advanced software engineering work. Anthropic says the model performs better on the hardest coding tasks, including jobs that previously required close supervision. In practice, the company says users have reported being able to hand off complex programming work with more confidence because the model can handle long-running tasks with greater rigor, follow instructions more precisely, and check its own outputs before responding.
A stronger push into coding and professional work
Anthropic is positioning Opus 4.7 as a model that is not just smarter in the abstract, but more useful for demanding day-to-day work. The company says it is especially effective when the task involves extended planning and execution, where consistency matters as much as raw speed. It also says the model is better at catching logical flaws during the planning phase, which can reduce rework later in the process.
The company’s early-access testers echoed that emphasis. One tester from a financial technology platform said Opus 4.7 catches its own logical faults during planning and accelerates execution, describing the combination of speed and precision as potentially “game-changing” for development velocity. Another early evaluator from Anthropic’s materials said Claude Opus 4.7 pushes the company’s coding performance further and stands out for more than raw capability, though the provided excerpt cuts off before completing that thought.
Anthropic also says the model has substantially improved vision capabilities. It can process images at higher resolution, and the company says it is more tasteful and creative when generating professional materials such as interfaces, slides, and documents. That suggests Anthropic is aiming Opus 4.7 at a broader set of workplace tasks, not only software engineering.
Cybersecurity safeguards come first
The release comes a week after Anthropic announced Project Glasswing, which focused on the risks and benefits of powerful AI models in cybersecurity. At that time, the company said it would keep Claude Mythos Preview’s release limited and test new cyber safeguards on less capable models before widening access.
Opus 4.7 is the first model to go through that process. Anthropic says its cyber capabilities are not as advanced as those in Mythos Preview, and that during training it experimented with ways to reduce those capabilities further. The company is releasing the model with safeguards that automatically detect and block requests tied to prohibited or high-risk cybersecurity uses.
Anthropic says the real-world deployment of those safeguards will help inform its eventual goal of a broader release for Mythos-class models. Security professionals who want to use Opus 4.7 for legitimate cybersecurity work, including vulnerability research, penetration testing, and red-teaming, are being invited to join a new Cyber Verification Program.
Available now at the same price
Protect your privacy with Doppler VPN
3-day free trial. No registration. No logs.
Claude Opus 4.7 is available today across Claude products, Anthropic’s API, Amazon Bedrock, Google Cloud’s Vertex AI, and Microsoft Foundry. Pricing remains unchanged from Opus 4.6 at $5 per million input tokens and $25 per million output tokens. Developers can access it through the Claude API using claude-opus-4-7.
The release underscores Anthropic’s strategy of pairing capability gains with tighter controls, especially as frontier models become more useful in sensitive domains. By improving performance in coding, image understanding, and professional content generation while simultaneously testing cybersecurity guardrails, Anthropic is trying to balance broader utility with narrower, more deliberate access to higher-risk functions.
For now, Opus 4.7 appears to be Anthropic’s strongest public offering for users who want a model that can take on difficult technical work with less oversight than before, while the company continues to keep its most capable cyber-related systems on a shorter leash.
Sources:
Read more tech news on the Doppler VPN Blog.