Anthropic releases Claude Opus 4.7 with stronger software and vision capabilities

After a week that has barely given anyone room to breathe, Anthropic is back in the spotlight with the launch of Claude Opus 4.7, its most powerful generally available model so far. The release comes soon after Mythos Preview, Anthropic’s more powerful but privately limited cybersecurity model, and is positioned as a direct upgrade over Opus 4.6, with better performance in advanced software engineering, complex coding tasks, and multi step development workflows, while also following instructions more reliably.

The model also improves consistency on long running tasks, reduces the need for close user guidance, and brings upgrades to vision by supporting higher resolution image analysis. Anthropic says Opus 4.7 is also better at creative tasks such as producing slides and documents, and that it now leads agentic coding benchmarks again, outperforming OpenAI’s GPT 5.4.

At the same time, Anthropic says Opus 4.7 is intentionally less capable in cybersecurity than Mythos Preview, which remains restricted to selected partners including Nvidia, JPMorgan Chase, Google, Apple, and Microsoft due to security concerns. The company is using Opus 4.7 to test new cybersecurity safeguards that could help inform a broader release of Mythos class models later on. Anthropic is also launching a Cyber Verification Program for security researchers who want to use Opus 4.7 for tasks such as vulnerability research, potentially with fewer restrictions than the standard release. Opus 4.7 is available today across all Claude products, the Claude API, Amazon Bedrock, Google Cloud Vertex AI, and Microsoft Foundry, with pricing unchanged from Opus 4.6.