Compute Screener
Updated 8:24 PM CST

Claude Opus 4.8 API price per 1M tokens: input, cached and output prices, price chart and history

# Lab/Bloomberg ticker USD / 1M tokens 1D 1W 1M Trend Chart (3M)
1$0.613+4.1%+3.4%-27.4%▼ Strong 7.6-20.6% a month
2$0.0636-18.5%-0.2%-42.4%▼ Strong 6.9-54.2% a month
3$0.0276-0.7%-2.8%-3.8%▼ Strong 5.1-29.4% a month
4$0.224+1.3%+1.4%+218.6%▲ Strong 4.5+62.8% a month
5$0.0353+3.2%-0.3%+32.7%▲ Moderate 3.6+13.0% a month
6$0.121-6.4%-29.0%-25.1%▼ Moderate 3.1-30.3% a month
7$0.0432-0.2%+3.6%-56.9%▼ Moderate 2.7-20.1% a month
8$0.236+4.2%+11.8%+68.7%▲ Moderate 2.2+12.4% a month
9$0.7350.0%0.0%0.0%▲ Moderate 2.1+4.6% a month
10$0.01570.0%0.0%0.0%▼ Weak 1.8-47.1% a month
11$0.0404-79.8%-79.8%-79.8%▼ Weak 1.7-20.4% a month
12$0.161+7.0%-13.9%-48.2%▼ Weak 1.4-11.9% a month
13$0.628+7.8%-5.3%-18.6%▼ Weak 1.3-2.8% a month

Claude Opus 4.8Anthropic

$1.290.0% (All)
USD / 1M tokens
Cache read $0.50 · Input $5.00 · Output $25.00 per 1M tokens, at 87% cache read, 12% input, 1% output
1D
1W
1M
3M
6M
YTD
1Y
—
—
—
0.0%
—
—
—
#ModelAs ofShareUSD / 1M tokens
1Claude Opus 5.5Nov 3, 8:25 PM CST53.6%$0.854
2Claude Haiku 5.5Nov 3, 8:25 PM CST19.3%$0.0257
3Claude Sonnet 5.5Nov 3, 8:25 PM CST18.8%$0.427
4Claude Sonnet 5Nov 3, 8:25 PM CST4.7%$0.514
5Claude Opus 4.8Nov 3, 8:25 PM CST3.6%$1.29

Break-even throughput by GPU

GPUUSD / GPU-hourTokens/s per GPU
B300$9.06252
B200$7.85218
H200$5.10142
H100$3.70103
A100$1.7549
L40S$1.4039
L40$0.9627
RTX A6000$0.5816
RTX 4090$0.5315

Gross margin calculator

Revenue—
GPU rental price$3.70
Gross profit—
Gross margin—

About Claude Opus 4.8

Claude Opus 4.8 is a model from Anthropic. Its posted API price is $5.00 per 1M input tokens, $0.50 per 1M cached input tokens and $25.00 per 1M output tokens, or $1.29 per 1M tokens at the mix of tokens agents send (87% cache read, 12% input, 1% output). Its price has not changed since May 28, 2026. It carries 3.6% of Anthropic's API tokens and is one of the 5 models in the Anthropic token price index, $0.613 per 1M tokens today. Of the 36 models in the token index it is the most expensive. At today's H100 rental price of $3.70 per GPU-hour, its break-even throughput is 103 tokens per second per GPU at its list price (3 input : 1 output). Prices are collected once a day, for each UTC day.

Models near this price

  • GPT-5.6 Sol (OpenAI): $1.03 per 1M tokens at the mix, $4.00 input and $20.00 output
  • Claude Opus 5.5 (Anthropic): $0.854 per 1M tokens at the mix, $4.00 input and $20.00 output
  • Kimi K3 (Moonshot AI): $0.771 per 1M tokens at the mix, $3.00 input and $15.00 output
  • Grok 4.7 (xAI): $0.735 per 1M tokens at the mix, $2.00 input and $6.00 output
  • Claude Sonnet 5 (Anthropic): $0.514 per 1M tokens at the mix, $2.00 input and $10.00 output
  • Kimi K3.1 Pro (Moonshot AI): $0.514 per 1M tokens at the mix, $2.00 input and $10.00 output

Cite as: Compute Screener, “Claude Opus 4.8 API price”, computescreener.com/models/claude-opus-4-8. The figures are free to reuse under CC BY 4.0 with credit to Compute Screener.