Claude Opus 4.8 is our most intelligent Opus model and the best generally available model for coding and agents, with deeper reasoning for enterprise workflows.

Claude Opus 4.8 is our most intelligent Opus model and the best generally available model for coding and agents, with deeper reasoning for enterprise workflows.

Model Details

Details are as follows

This model supports the following features:

• Memory tool (beta): Store and retrieve information across conversations for enhanced context management
• 1M token context window: Process and maintain coherent understanding across extended documents and conversations
• Enhanced tool orchestration: Parallel tool execution and automatic tool call management
• Prompt caching
• Batch prediction
• Count tokens
• Global endpoint
• Multi-region endpoint
• Note: Opus 4.8’s quota is included in the new Opus-family quota, which provides a single shared quota for all Opus models going forward.

• Coding: Claude Opus 4.8 is the best coding model for real-world work, and runs independently for longer than previous Opus models. It reads codebases like a subject matter expert and plans before it edits. On long, multi-stage tasks it keeps track of dependencies, gets unstuck on its own, and persists over hours with stronger memory and longer context.
• Enterprise workflows: Claude Opus 4.8 is built for scalable workflows. It is better at following instructions, staying in scope, and producing professional-grade outputs. It reasons across long documents, checks its work, and manages multi-stage projects end-to-end with professional polish on spreadsheets, slides, and docs.
• Long-running agents: Claude Opus 4.8 sets a new bar for agentic workflows, pairing improved tool use with creative problem-solving across multi-step tasks. It is designed to work independently for longer, unblocking itself across multi-stage projects and complex dependency chains with minimal oversight.
• Financial analysis: Claude Opus 4.8 brings deeper reasoning and precision to financial workflows like investment research and earnings analysis, reading dense filings and charts at high fidelity and carrying context across an entire reporting cycle.
• Cybersecurity: Claude Opus 4.8 brings deeper reasoning to security workflows like threat intelligence synthesis, vulnerability finding, alert triage, and incident response. It holds long traces and large codebases in context to help catch subtle patterns and complex attack vectors.
• Computer use: Claude Opus 4.8 is our most capable generally available computer-use model, pairing improved vision with deep reasoning for multi-step tasks that span multiple applications and require planning and judgment.