Anthropic Launches Claude Opus 4.7 AI Model With Enhanced Coding and Visual Features

Anthropic has released its latest AI model, Claude Opus 4.7, making it available to all users. This launch follows the recent announcement of Claude Mythos, which remains exclusive to about 40 companies due to its advanced capabilities. Opus 4.7 is described as Anthropic’s most powerful model for general users, offering significant improvements over the previous version, Claude Opus 4.6.

Key Highlights

• Anthropic launches Claude Opus 4.7 for all users with improved coding and visual reasoning
• Opus 4.7 processes high-resolution images up to 2,576 pixels and handles complex software tasks
• Model consumes more tokens, prompting Anthropic to increase rate limits for all subscribers
• Claude Opus 4.7 includes enhanced safeguards for cybersecurity and better instruction adherence
• Pricing remains unchanged at $5 per million input tokens and $25 per million output tokens

Key Features of Claude Opus 4.7

Claude Opus 4.7 excels in software engineering, instruction adherence, and real-world task completion. The model can handle complex coding assignments that previously required close human supervision. Anthropic has also improved the model’s ability to process high-resolution images up to 2,576 pixels. This enhancement allows Opus 4.7 to better interpret dense screenshots and complex diagrams, which older versions struggled to analyze.

The model consumes more tokens, which are units of AI processing. To address this, Anthropic has increased token rate limits for all subscribers. Boris Cherny, head of Claude Code, confirmed this change on X, stating that the adjustment ensures users do not run out of their token limits when using Opus 4.7.

Comparison With Other Anthropic Models

While Opus 4.7 is a major step up from Opus 4.6, it is still "less broadly capable" than Claude Mythos Preview. Mythos is Anthropic’s most advanced model, especially in cybersecurity, but is restricted to select companies such as Microsoft, Google, and Amazon under Project Glass wing. Anthropic limited Mythos’s availability due to concerns over its ability to identify cybersecurity vulnerabilities, which could be misused.

Opus 4.7, however, includes more safeguards than Mythos. It can automatically block high-risk cybersecurity requests, reducing the risk of misuse. The model also demonstrates improved instruction following, offering more precise responses to user prompts compared to earlier versions.

Availability and Pricing

Claude Opus 4.7 is now accessible through all Claude products and cloud platforms, including Amazon Bedrock, Google Cloud’s Vertex AI, and Microsoft Foundry. Anthropic has kept the pricing unchanged from Opus 4.6, charging $5 (about Rs 465) per million input tokens and $25 (about Rs 2,325) per million output tokens.

Anthropic states that Opus 4.7 is a distinct model, not related to Mythos, and serves as a successor to Opus 4.6. The company positions Opus 4.7 between its earlier models and the more advanced Mythos in terms of capability and access.