★ FeaturedFlagship — best answers, highest costCommercial API

Claude Opus 4.8 Pricing

Anthropic · claude-opus-4-8

Input / 1M tokens

$5.00

Output / 1M tokens

$25.00

Context window

1M tokens

≈ 1,333 pages of text

Max output

128k tokens

What it costs in practice

Typical request (1,200 in + 400 out tokens)$0.016

1,000 requests / month$16.00/mo 10,000 requests / month$160.00/mo 100,000 requests / month$1,600.00/mo

Estimate your own workloadOpens the calculator with Claude Opus 4.8 preloaded — adjust volume and token counts there.

Price history

Only one price point recorded so far.

The staircase chart appears once a price change is detected.

What it excels at

Best-in-class for complex multi-step agentic workflows, long-horizon planning, and large codebase refactors. Strong instruction-following on ambiguous, open-ended tasks.

The business tradeoff

Highest per-token cost in the catalog and noticeably higher latency — reserve for tasks that genuinely need top-tier reasoning rather than routine completions.

Cheaper from Anthropic

Claude Sonnet 4.5Anthropic$0.0096/ typical request Claude Sonnet 4.6Anthropic$0.0096/ typical request Claude Haiku 4.5Anthropic$0.0032/ typical request

Same tier from other providers

GPT Chat LatestOpenAI$0.018/ typical request Grok 4.3xAI$0.0025/ typical request Gemini 3.1 Pro Preview Custom ToolsGoogle$0.0072/ typical request

Head-to-head

Claude Opus 4.8 vs GPT Chat Latest →Claude Opus 4.8 vs Grok 4.3 →Claude Opus 4.8 vs Gemini 3.1 Pro Preview Custom Tools →

Vendor list rates, as of Jun 15, 2026 · source: openrouter · per-request examples assume 1,200 input + 400 output tokens.