Models & pricing / claude-opus-4-8 · Anthropic
Anthropic

claude-opus-4-8 API

Claude Opus 4.8 builds on 4.7 for coding, agent collaboration and knowledge work, with text and image input and text output. It remains in Anthropic’s active legacy lineup. This XiuAI page provides current protocols, pricing and migration references for existing integrations.[1][2][4]

claude-opus-4-8

Model overview

Published provider specifications. Available features depend on the current API route.

API routes
OpenAI-compatible API / openai-response / openai-response-compact / Anthropic Messages / Gemini GenerateContent / openai-alpha-search

Multi-step collaboration and analysis

Consider it for engineering investigations, cross-file revisions and research that require repeated context checks. Anthropic’s release emphasizes collaboration across tool use, reasoning and knowledge work. Define intermediate outputs and failure conditions in your application so each step can be checked.[2]

Release-page testimonials cover legal, financial and browser-agent work, but they describe specific customer settings rather than universal accuracy or suitability. Preserve original evidence, citation locations and human review where professional judgment is required.[2]

What changes when moving from 4.8

If an older Opus 4.8 request did not enable thinking, moving to always-thinking Opus 5.5 can change usage and content-block structure. Recheck max_tokens, select text by type, and ensure your application does not trim or reorder thinking blocks that must be returned.[3]

API access

Access, charges and runtime conditions

Use claude-opus-4-8 with your existing XiuAI key and a native or compatibility protocol listed in the catalog. Version maintenance, migration and live availability are separate questions; legacy status does not mean XiuAI has stopped serving the model.[1][4]

Costs follow current input, output and cache conditions. The launch also introduced Claude Code dynamic workflows and vendor fast mode. Those are product-specific features or billing modes, not workflows or free acceleration promised by XiuAI.[2][4]

Send a request using your existing API key.

POST /v1/chat/completions
curl 'https://router-api.xiu.ai/v1/chat/completions' \
  -H "Authorization: Bearer $XIUAI_API_KEY" \
  -H 'Content-Type: application/json' \
  --data-raw '{
  "model": "claude-opus-4-8",
  "messages": [
    {
      "role": "user",
      "content": "Hello"
    }
  ]
}'
Read the API documentation

Frequently asked questions

Does Opus 4.8 include a complete browser workspace?

No. The model can contribute tool calls and visual understanding, but the browser, permissions, execution and file storage belong to the application. Browser-agent examples on the launch page do not establish a XiuAI product feature.[2][4]

Why can thinking blocks appear first after moving to 5.5?

Opus 5.5 always enables thinking, while an older 4.8 request may not have used it. Handle content blocks by type instead of assuming the first block contains text. Thinking and text also share the output budget.[3]

Can I budget using performance or cost gains from the launch page?

First check the evaluation setting and request conditions. Budget using current XiuAI rates and your actual usage; testimonials, vendor benchmarks and fast-mode conditions do not replace measurements for your integration.[2][4]

Related models

claude-opus-5claude-opus-5-5

Sources and updates

Specifications reference provider documentation; prices and availability come from the current XiuAI catalog.

Content reviewed: 2026-11-16

[1] Claude Opus 4.8 model reference[2] Claude Opus 4.8 announcement[3] Anthropic migration guide[4] Current XiuAI model and pricing dataModel documentationTool callingStructured outputContext documentation