Claude Opus 5.5 pricing is $4.00/1M input and $20.00/1M output per 1M tokens. Cached input is $0.200/1M (5% of input). Cache writes are $5.00/1M for the 5-minute catalog line. About 20% cheaper than Opus 5 on stickers, with much cheaper cache reads. Rates in TokenCalculator checked 23 Sep 2026.
Price a Claude workload
In-browser Claude counts are Approx. Use the cost calculator for planning, then confirm critical budgets with Anthropic’s count_tokens API.
Claude Opus 5.5 rate card
API model id: claude-opus-5-5. Context window: 1M tokens at standard rates across the window. Max output: 128K. Anthropic publishes a separate one-hour cache write tier ($8 / 1M) in official docs. TokenCalculator stores the 5-minute write line when one cache write field is available.
| Line | USD per 1M tokens |
|---|---|
| Input | $4.00/1M |
| Cached input (hits / refreshes) | $0.200/1M |
| Cache write (5m, catalog) | $5.00/1M |
| Output | $20.00/1M |
| Batch input | $2.00/1M |
| Batch output | $10.00/1M |
Cheaper than Opus 5, still below Fable
| Model | Input / 1M | Output / 1M | Cache read / 1M |
|---|---|---|---|
| Claude Opus 5.5 | $4.00/1M | $20.00/1M | $0.200/1M |
| Claude Opus 5 | $5.00/1M | $25.00/1M | $0.500/1M |
| Claude Fable 5.1 | $10.00/1M | $50.00/1M | $0.250/1M |
| GPT-6 Sol | $2.00/1M | $10.00/1M | $0.200/1M |
Anthropic positions Opus 5.5 as the default Opus for complex agentic coding and enterprise work. Prefer Fable 5.1 when Opus still fails long-horizon evals. Prefer Sol when OpenAI Exact counts and lower stickers win your harness.
Cache reads and Fast mode
Cache hits and refreshes are $0.20 per 1M. Five-minute writes are $5 per 1M. One-hour writes are $8 per 1M on Anthropic docs. Batch is 50% of standard. Fast mode on the Claude API is $8 / $40 (about 2x). Agent loops that reread a stable prefix feel the cache read line more than the sticker.
No OpenAI-style 272K cliff
Opus 5.5 publishes standard per-token pricing across its 1M window. There is no Astra or Sol long-context band that doubles rates past 272K. Million-token windows are still not free. You still pay for every token, and Approx browser counts need count_tokens before contracts.
Worked examples
- Fresh 10K input + 2K output about $0.08 on standard rates.
- Same turn with 8K of that input cached: cache read $0.0016 + 2K fresh $0.008 + 2K output $0.04 about $0.05.
- Batch halves input and output when overnight latency is fine.
Tokenizer note
Do not budget Opus 5.5 with tiktoken or OpenAI Exact counts. TokenCalculator labels Claude Approx. Confirm shipping budgets with Anthropic count_tokens.
Common mistakes
- Assuming Opus 5.5 is always cheaper than Sol because cache reads match at $0.20 (Sol stickers are still half)
- Ignoring cache write cost on the first pass of a large prefix
- Using OpenAI Exact counts to budget Claude spend
- Picking Fable for routine work that Opus 5.5 would finish cheaper
- Turning on Fast mode without checking the 2x bill
Frequently asked questions
How much does Claude Opus 5.5 cost?
Standard Claude API pricing is $4.00/1M per 1M input and $20.00/1M per 1M output. Cache reads are $0.200/1M.
What is the Claude Opus 5.5 model id?
Use claude-opus-5-5 on the Claude API.
Is Opus 5.5 cheaper than Opus 5?
Yes on list rates in this catalog. Opus 5 is $5.00/1M / $25.00/1M with $0.500/1M cache reads. Opus 5.5 cuts stickers and cache reads.
Does Opus 5.5 have a long-context surcharge?
Anthropic documents standard per-token pricing across the 1M window. There is no OpenAI-style 272K cliff on the published Opus 5.5 card.
Do cache reads cost the same on Sol and Opus 5.5?
Yes on the catalog cache read line: both $0.20 per 1M. Cache writes and input/output stickers still differ.
When should I use Fable 5.1 instead?
When Opus 5.5 still fails long-horizon agents or hard research evals and you accept $10 / $50 list rates with cheaper Fable cache reads ($0.25).
Are browser token counts Exact for Claude?
No. TokenCalculator labels Claude Approx. Confirm shipping budgets with Anthropic count_tokens.
Are these rates live?
No. Curated list prices with lastVerified 23 Sep 2026. Confirm on Anthropic’s pricing page.
Next steps
- GPT-6 Sol vs Claude Opus 5.5
- GPT-6 Sol pricing
- Claude Fable 5.1 pricing
- Claude API pricing hub
- Prompt caching guide
- Official Anthropic pricing
Related Sep 2026 pricing guides
Frontier rate cards, cost-per-task math, and long-context cliffs. Cross-link these when you compare Sol, Luna, Opus 5.5, Grok 4.7, or Astra.