Claude Opus 5 API Pricing
$5.00 input · $25.00 output per 1M tokens — Anthropic's flagship for the hardest requests.
Claude Opus 5 is Anthropic's flagship: $5.00 input and $25.00 output per million tokens, with cached input at $0.50. It's the top-tier choice for the hard 10% of requests where mid-tier or budget models fall short — the other 90% of traffic is usually better served by Sonnet 5 or Haiku 4.5.
Price per 1M tokens
Claude Opus 5 · USD official list price
| Tier | Rate |
|---|---|
| Input | $5.00 |
| Output | $25.00 |
| Cache hit | $0.50 |
DeepSeek lists official prices in CNY; USD uses ≈ 6.9 exchange rate
Best for
$5.00 / $25.00
Flagship list price for input and output per million tokens.
$0.50 cached input
Cached input at one-tenth of the uncached rate.
50% Batch discount
Anthropic's Batch API halves the cost of offline workloads.
High-stakes, complex reasoning and generation where output quality is the binding constraint — and where Sonnet 5 or Haiku 4.5 demonstrably doesn't meet the bar.
Billing notes
Published API pricing as of September 2026, standard tier, USD per million tokens. Anthropic offers a 50% Batch API discount for offline workloads, bringing output to $12.50.
Opus 5 vs the competition
Against GPT-5.6 Sol ($5/$30), Opus 5 undercuts Sol's output by ~17% ($25 vs $30) at the same input price; against DeepSeek V4 Pro, it loses badly on price — V4 Pro's ¥4.50 off-peak output is a fraction of Opus 5's $25.00.
Frequently Asked Questions
What is the price of Claude Opus 5 per million tokens?
Claude Opus 5 bills $5.00 per million input tokens and $25.00 per million output tokens, with cached input at $0.50 per million. Anthropic offers a 50% Batch API discount for offline workloads.
When is Opus 5 worth the flagship price?
Opus 5 is the top-tier choice for the hard 10% of requests where mid-tier or budget models fall short. For the other 90% of traffic, a mid-tier or budget model is typically the better value.
Does Opus 5 have prompt caching?
Yes. Claude Opus 5 charges $0.50 per million tokens for cached input — one-tenth of the uncached rate. Keeping a stable system prompt and reusable context prefix meaningfully reduces input cost.
Related model pricing
Compare this model against the rest of the lineup.
Estimate your exact cost
Model your own workload — input/output tokens, cache hit rate, and peak-hour share — against any model mix with the interactive pricing calculator.
Open the Pricing CalculatorRecommended Tools We ARE USING
We are using these tools ourselves for Development / Deployment. Check out for more details.
Opencode Go
You will receive $5 when subscribed, which directly offsets the 1st month's Go fee.
Could be the cheapest option for calling DeepSeek V4 Flash model in the world.
Vultr
RReceive $300 to test out VVultr platform
Referred user must be active 30+ days and spend $10–$25
Railway
Deploy anything without the complexity
Connect your repo, Railway handles the rest.
腾讯云 Tencent Cloud
Cloud servers, cloud databases, COS, CDN, SMS and other cloud products are on special offer now.
AWS high-quality alternative, we use COS as a replacement for S3
硅基流动 SiliconFlow
¥16 universal voucher for all platforms, valid for 180 days from the date of receipt
A comprehensive product matrix supporting the full-process implementation of AI applications.
Warp
One of the best AI terminal tools in the world
Interestingly, we can directly use various top-tier models in Warp without leaving the terminal.