How Much Does Claude Opus Cost? Opus 5 and Every Opus Price
In short: Claude Opus costs $4 per million input tokens and $20 per million output tokens for the current model, Claude Opus 5.5, through Anthropic's API. Claude Opus 5, released on July 24, 2026, costs $5 and $25. In the Claude apps you pay a plan price instead: Free, Pro at $20 a month or Max from $100 a month. Every Opus model's price is in the table below, from Anthropic's pricing page.
How much does Claude Opus cost?
Each Opus model has its own price, set per million tokens of input (what you send) and output (what it writes). A token is roughly three-quarters of an English word.
| Opus model | Input (per 1M tokens) | Output (per 1M tokens) | Cache hit (per 1M tokens) |
|---|---|---|---|
| Claude Opus 5.5 (current) | $4 | $20 | $0.20 |
| Claude Opus 5 | $5 | $25 | $0.50 |
| Claude Opus 4.8 | $5 | $25 | $0.50 |
| Claude Opus 4.7 | $5 | $25 | $0.50 |
| Claude Opus 4.6 | $5 | $25 | $0.50 |
| Claude Opus 4.5 | $5 | $25 | $0.50 |
| Claude Opus 4.1 | $15 | $75 | $1.50 |
| Claude Opus 4 | $15 | $75 | $1.50 |
Source: Anthropic API pricing. A "cache hit" is the reduced price for input the model has recently seen, such as a long document sent again.
Worked example. Sending Claude Opus 5.5 a 20,000-token prompt (about 15,000 words) and getting a 2,000-token answer costs 20,000 × $4 ÷ 1,000,000 = $0.08 for input, plus 2,000 × $20 ÷ 1,000,000 = $0.04 for output: $0.12 in total. The same request on Opus 5 costs $0.15.
In the Claude apps (claude.ai, desktop and mobile), you do not pay per token. Anthropic's plans are:
- Free - $0.
- Pro - $20 a month, or $17 a month billed annually ($200 up front).
- Max - from $100 a month, with 5x or 20x more usage than Pro.
The pricing page lists "more Claude models" as a Pro feature but does not say which plan includes which model.
What is Claude Opus 5?
Claude Opus 5 is a large model from Anthropic for long, complex work such as agentic coding. "Agentic" means the model works through a task on its own - writing code, running tools, reading the results and continuing - rather than answering one message at a time.
In Anthropic's range, Opus sits between the faster Sonnet and the highest-capability Fable line. Anthropic's models overview describes the current Opus, Opus 5.5, as "for long-running agentic coding and knowledge work". Opus 5 now appears under "additional models" on the pricing page.
When was Claude Opus 5 released?
Claude Opus 5 was released on July 24, 2026. Claude Opus 5.5 followed on September 22, 2026, at a lower price ($4/$20 against $5/$25).
What are Claude Opus 5's specs?
- Context window: 1 million tokens - about 750,000 words of English at once.
- Maximum output: 128,000 tokens in one reply.
- Input: text and images; output: text.
- Thinking: adaptive thinking - it can reason step by step before answering.
- Tools: function calling, structured output, code execution and web search through the API.
- Long prompts: the whole 1-million-token window is billed at one rate, with no surcharge for long prompts.
Anthropic does not publish how many parameters Opus models have.
Claude Opus 5 vs Opus 5.5 vs Opus 4.8
| Claude Opus 5.5 | Claude Opus 5 | Claude Opus 4.8 | |
|---|---|---|---|
| Released | September 22, 2026 | July 24, 2026 | May 28, 2026 |
| Input / output price | $4 / $20 | $5 / $25 | $5 / $25 |
| Context window | 1M tokens | 1M tokens | 1M |
| Max output | 128K tokens | 128K tokens | 128K tokens |
| LMArena text | 1509 Elo (rank 3, high effort) | 1492 Elo (rank 14, high effort) | 1474 Elo (rank 41) |
Switching between Opus versions in the API is a change of model ID (for example claude-opus-5-5), not a different interface. Compare them side by side.
How does Claude Opus rank?
On LMArena's text leaderboard, which ranks models by blind human votes, Claude Opus 5.5 has 1509 Elo (rank 3, high effort) and Claude Opus 5 has 1492 Elo (rank 14, high effort). See the live order on our benchmarks table.
Related
Frequently asked questions
How much does Claude Opus cost?+
Through Anthropic's API, the current model, Claude Opus 5.5, costs $4 per million input tokens and $20 per million output tokens. Claude Opus 5, 4.8, 4.7, 4.6 and 4.5 cost $5 and $25, and Opus 4.1 and Opus 4 cost $15 and $75. In the Claude apps, plans are Free, Pro at $20 a month and Max from $100 a month.
How much is Claude Opus per month?+
Anthropic does not sell Opus on its own as a monthly subscription. Its Claude app plans are Free, Pro at $20 a month or $17 a month billed annually, and Max from $100 a month; the pricing page does not say which plan includes which model. Developers pay per token through the API instead.
What is Claude Opus 5?+
Claude Opus 5 is a large language model from Anthropic for long, complex tasks such as agentic coding. It has a 1-million-token context window and costs $5 per million input tokens and $25 per million output tokens through the API.
When was Claude Opus 5 released?+
Claude Opus 5 was released on July 24, 2026. Its successor, Claude Opus 5.5, was released on September 22, 2026.
Is Claude Opus 5.5 cheaper than Opus 5?+
Yes. Claude Opus 5.5 costs $4 per million input tokens and $20 per million output tokens, against $5 and $25 for Opus 5, and its cache-hit price is $0.20 against $0.50.
What is the difference between Claude Opus and Claude Sonnet?+
Opus is Anthropic's larger model for long, complex work, and Sonnet is the mid-sized model balancing speed and capability. Claude Opus 5.5 costs $4 and $20 per million input and output tokens; Claude Sonnet 5.5 costs $2 and $10.
Related Guides
Enjoyed this? Get more in your inbox.
Up to ten of the week's AI news stories, by email once a week.