Claude Opus 4.8 on Google Cloud

Claude Opus 4.8 on Google Cloud is a high-intelligence Opus model built for coding and agents, featuring deep reasoning for enterprise workflows.

Retirement Date: Not sooner than May 28, 2027.

  • Coding: Claude Opus 4.8 on Google Cloud is an advanced 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 on Google Cloud 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 on Google Cloud 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 on Google Cloud 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 on Google Cloud 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 on Google Cloud is a highly capable model for computer-use tasks, pairing improved vision with deep reasoning for multi-step tasks that span multiple applications and require planning and judgment.

Try in Agent Studio View model card in Model Garden

Model ID claude-opus-4-8
Launch stage GA
Supported inputs & outputs
  • Inputs:
    Text, Image, PDF
  • Outputs:
    Text
Token limits
  • Maximum input tokens: 1,000,000
  • Maximum output tokens: 128,000
Capabilities
Usage types
Technical specifications
Images
  • Limitation and specifications: See Vision in Anthropic's documentation
Documents
  • Limitation and specifications: See PDF support in Anthropic's documentation
Versions
  • claude-opus-4-8
    • Launch stage: Generally available
    • Release date: May 28, 2026
Supported regions

Model availability

(Includes fixed quota & Provisioned Throughput)

  • United States
    • Multi-region
  • Europe
    • Multi-region
  • Global
    • global endpoint

ML processing

  • United States
    • Multi-region
  • Europe
    • Multi-region
Quota limits

Multi-region:

Multi-region:

global endpoint:

Pricing See Pricing.