Claude Haiku 4.5: Pricing, Context, and Benchmarks
Claude Haiku 4.5 is Anthropic's Claude 4 model released on 2025-10-01. It runs on a 200K-token context window, priced at $1.00/M input and $5.00/M output. Fastest Claude tier. Good for routing, classification, and any latency-sensitive task.
Specifications
| Provider | Anthropic |
|---|---|
| Family | Claude 4 |
| Released | 2025-10-01 |
| Status | active |
| Context window | 200K tokens |
| Max output | 16K tokens |
| Modalities | text image |
| Capabilities | function calling json mode vision streaming tool use |
| Tokenizer | claude-3 |
| API endpoint | https://api.anthropic.com/v1/messages |
| Documentation | Official docs ↗ |
| Last updated | 2026-05-29 |
What this model is good for
Based on benchmark scores and capability fit, Claude Haiku 4.5 ranks well for:
- Customer Support — rank #2. Faster than Sonnet, similar instruction following quality.
Cost calculator
For 1,000 tokens of input and 1,000 tokens of output per call, this model costs $0.0060 per call. For typical usage:
| Workload | Tokens / call (in/out) | $/call | $/1K calls | $/1M calls |
|---|---|---|---|---|
| Short Q&A | 200 / 100 | $0.000700 | $0.70 | $700 |
| Standard chat | 1K / 500 | $0.003500 | $3.50 | $3,500 |
| RAG with retrieval | 4K / 500 | $0.006500 | $6.50 | $6,500 |
| Long doc summary | 20K / 1K | $0.025000 | $25.00 | $25,000 |
| Long context (100K input) | 100K / 2K | $0.107500 | $107.50 | $107,500 |
For more precise estimates including caching discounts, use the LLM API Cost Calculator.
Other Anthropic models
Common comparisons
- Claude Haiku 4.5 vs o3-mini Side-by-side specs, pricing, and benchmarks
- Claude Haiku 4.5 vs GPT-5 Side-by-side specs, pricing, and benchmarks
- Claude Haiku 4.5 vs Gemini 2.5 Pro Side-by-side specs, pricing, and benchmarks
- Claude Haiku 4.5 vs Gemini 1.5 Pro Side-by-side specs, pricing, and benchmarks
API setup
curl https://api.anthropic.com/v1/messages \
-H "Authorization: Bearer $API_KEY" \
-H "Content-Type: application/json" \
-d '{"model": "claude-haiku-4-5", "messages": [{"role": "user", "content": "Hello"}]}'See the official documentation for the full request/response schema. For pre-flight token counting before paying, use the appropriate token counter.