Claude Sonnet 5.5 vs Claude Haiku 4.5
Claude Sonnet 5.5 (Anthropic) costs $2.00 per million input tokens and $10.00 per million output tokens; Claude Haiku 4.5 (Anthropic) costs $1.00 and $5.00. Claude Haiku 4.5 is 2.0× cheaper for a typical 3:1 input-to-output mix. Claude Sonnet 5.5 has the larger context window (1M vs 200K tokens).
| Claude Sonnet 5.5 | Claude Haiku 4.5 | |
|---|---|---|
| Provider | Anthropic | Anthropic |
| Input / 1M tokens | $2.00 | $1.00 |
| Output / 1M tokens | $10.00 | $5.00 |
| Cached input / 1M | $0.20 | $0.10 |
| Blended (3:1) | $4.00 | $2.00 |
| Context window | 1M | 200K |
| Max output | 128K | 64K |
| Vision | Yes | Yes |
| Tool use | Yes | Yes |
| Reasoning | Yes | Yes |
| Listed since | Oct 1, 2026 | Oct 18, 2025 |
Cost per 1,000 requests
| Workload | Claude Sonnet 5.5 | Claude Haiku 4.5 |
|---|---|---|
| Chatbot reply2,000 input + 500 output tokens | $9.00 | $4.50 |
| RAG / search answer8,000 input + 700 output tokens | $23.00 | $11.50 |
| Long-document summary100,000 input + 2,000 output tokens | $220.00 | $110.00 |
Blended price history
Blended = (3 × input + output) ÷ 4, per million tokens. Each line starts when the model first appeared in the price list.
The companies
List prices for each provider's own API (USD), taken daily from LiteLLM's public model price list. Batch, caching and committed-use discounts aren't reflected. Price alone says nothing about output quality; check benchmarks and your own evaluations before switching.