Claude Haiku 5.5
- Input / 1M tokens
- $0.10
- Output / 1M tokens
- $0.50
API models · Side by side
Claude Haiku 5.5 vs Claude Opus 5.5
Haiku 5.5 and Opus 5.5 sit at opposite ends of Anthropic's range. Sorting thousands of support emails and fixing a tricky bug are very different jobs, and these two models are priced for those different jobs.
Claude Opus 5.5
- Input / 1M tokens
- $4.00
- Output / 1M tokens
- $20.00
The short answer
Haiku charges $0.10 per million input tokens and $0.50 per million output tokens; Opus charges $4 and $20, 40 times more. Both handle text, images and files, read up to 1 million tokens and write up to 128,000 per reply. Haiku's rates rise to $0.50 and $2.50 above 100,000 tokens, which is still far below Opus.
Compare the details
Compare context, supported inputs and usage costs. Highlighted rows show a difference.
| Feature | Claude Haiku 5.5 | Claude Opus 5.5 |
|---|---|---|
| Input formats | file, image, text | file, image, text |
| Output formats | text | text |
| Context / position limit | 1,000,000 | 1,000,000 |
| Maximum output tokens | 128,000 | 128,000 |
| Developer features | Reasoning output, Completion limit, Output limit, Reasoning controls, Thinking effort, Response format, Stop sequences, Structured output, Tool selection, Tool calling, Response length | Reasoning output, Completion limit, Output limit, Reasoning controls, Thinking effort, Response format, Stop sequences, Structured output, Tool selection, Tool calling, Response length |
| Cached input / 1M tokens | $0.01 | $0.20 |
| Cache write / 1M tokens | $0.13 | $5.00 |
| USD input / 1M tokens | $0.10 | $4.00 |
| USD output / 1M tokens | $0.50 | $20.00 |
| Model identifier | anthropic/claude-haiku-5.5 | anthropic/claude-opus-5.5 |
API pricing
Prices are in USD per million tokens. Input is what you send; output is what the model generates. Claude and ChatGPT subscriptions are billed separately.
Claude Haiku 5.5: long-prompt rates
- Long-prompt threshold: 100,000 prompt tokens. Input: $0.50 / 1M. Output: $2.50 / 1M.
Tools, images and additional reasoning can add charges. The examples below cover input and output tokens only.
Compare local modelsWhat those prices mean in practice
Each example uses the same token budget for both models. These are calculations, not measured workloads. Cache discounts, media, tools and extra reasoning are excluded.
| Workload | Claude Haiku 5.5 | Claude Opus 5.5 |
|---|---|---|
| 1,000 short requestsPer request: 2,000 input + 500 output tokens | $0.45 | $18.00 |
| 100 document summariesPer request: 20,000 input + 1,000 output tokens | $0.25 | $10.00 |
| 10 long-document requestsPer request: 300,000 input + 2,000 output tokens | $1.55 | $12.40 |
Estimate = requests × (input tokens × input price + output tokens × output price) ÷ 1,000,000. Long-prompt rates are applied where the pricing data provides them. Real requests can use different amounts of output.
Beyond the price tag
Everyday work: Haiku
Sorting tickets, pulling fields from invoices and short summaries are ideal Haiku jobs. Give it a fixed answer format and clear rules for missing information. For example, leave the total empty if an invoice doesn't show one. Check answers automatically before they're used.
Hard problems: try Opus
For a difficult code change, or a report drawing on many documents, test Opus against Haiku on real examples. Judge the finished result: does it meet the requirements, and how much fixing is left? Opus costs about $18 per 1,000 short requests against Haiku's $0.45.
Use both together
Let Haiku sort incoming work, and send only the hard cases to Opus. Set a clear rule for the handover, such as missing information or a request outside the usual categories. Pass the original documents along, not just Haiku's summary.
Which one should you pick?
Choose Claude Haiku 5.5 for high-volume, well-defined tasks. Choose Claude Opus 5.5 only for complex work where your tests show it makes a real difference. Most teams get the best value by combining the two.
Common questions
Can Haiku replace Opus for everything?
No. Haiku is great for simple, checkable tasks. Complex work needs testing on both.
Are these Claude subscription prices?
No. They're pay-per-use API prices. The Claude app has separate plans.
Same context window, same answers?
No. Context size only says how much the model can read. It doesn't make the answers equal.