Back to home
AnthropicReleased 2026-05-29

Claude Opus 4.8

Anthropic's flagship model released May 29, 2026. Best-in-class for code, long-form analysis, and nuanced reasoning. $5/$25 per 1M tokens. The original Claude 4 series was retired June 15, 2026.

Context
1M tokens
Input
$5/1M
Output
$25/1M
Modalities
textimagePDF

Best For

Complex code refactoringLong document analysisNuanced writingHard debugging

Strengths

  • Best code reasoning in production models
  • Excellent at nuanced prose and brand voice
  • 1M context window + PDF understanding
  • Constitutional AI: more thoughtful responses
  • Long document analysis (PDFs, papers, books)

Weaknesses

  • More expensive than GPT-5.6 ($5 vs $2.50 input)
  • Slower than Flash variants
  • Stricter refusals (Microsoft CEO criticized this)
  • No native audio/video

Related Alternatives

GPT-5.6

Better for math/agents, half the price

Kimi K3

Better for Chinese, 5x cheaper

Frequently Asked Questions

How much does Claude Opus 4.8 cost?

Claude Opus 4.8 costs $5 per 1M input tokens and $25 per 1M output tokens. For Claude Pro subscription, it's $20/month for web access. Opus 4.8 Fast Mode (beta) costs $10/$50 per 1M for 2.5x faster responses.

Is Claude Opus 4.8 better than GPT-5.6 for coding?

Yes, generally. Claude Opus 4.8 leads GPT-5.6 on most coding benchmarks: SWE-Bench (78% vs 71%), multi-file refactoring, and code review accuracy. For simple code completion, GPT-5.6 Luna is faster and cheaper.

See Claude Opus 4.8's full record in our database

Live pricing for all Anthropic plans, token cost calculator, and the latest price changes.

View Full Comparison