Claude Haiku 5.5
Anthropic 最快、最省成本的模型，面向低延迟助手与大规模自动化。
上下文窗口
1M tokens
claude-haiku-5-5
最大输出
-
1M tokens
输入
$0.09$0.10
10% off1M tokens
输出
$0.45$0.50
10% off1M tokens
claude-haiku-5-5试玩区
claude-haiku-5-5暂无输出
定价
| 方案 | 上下文窗口 | 最大输出 | 输入 | 输出 | 缓存写入 | 缓存读取 |
|---|---|---|---|---|---|---|
claude-haiku-5-5 | 1M tokens | - | $0.09$0.1010% off 1M tokens | $0.45$0.5010% off 1M tokens | $0.1125$0.12510% off 1M tokens | $0.009$0.0110% off 1M tokens |
Long Context | - | - | $0.45$0.5010% off 1M tokens | $2.25$2.5010% off 1M tokens | $0.5625$0.62510% off 1M tokens | $0.045$0.0510% off 1M tokens |
Cache Write 1h · $0.18 · 1M tokens
Long Context Cache Write 1h · $0.9 · 1M tokens
Applies to the full request when the prompt exceeds 100,000 tokens
缓存读取或缓存输入价格适用于复用的提示词 tokens。
README
- Claude Haiku 5.5 专为低延迟助手、客服系统与大规模自动化打造，适用于对成本与响应时间敏感的场景。
- 它兼具快速、省成本的文字生成能力与提示缓存，可降低重复上下文成本。
- ReachAPI 以 `claude-haiku-5-5` 模型代码通过 Claude Messages 协议提供此模型。
API
API
端点
POST /v1/messages认证
Authorization: Bearer YOUR_REACH_API_KEY anthropic-version: 2023-06-01 Content-Type: application/json
示例
curl -X POST "https://direct.reachapi.ai/v1/messages" \
-H "Authorization: Bearer YOUR_REACH_API_KEY" \
-H "anthropic-version: 2023-06-01" \
-H "Content-Type: application/json" \
-d '{
"model": "YOUR_MODEL_ID",
"messages": [
{
"role": "user",
"content": "Introduce ReachAPI in one sentence."
}
],
"max_tokens": 512,
"temperature": 0.2
}'相关模型
Claude Opus 5.5
Anthropic's flagship model for deep reasoning, complex agentic coding, and long-horizon enterprise work.
Claude Sonnet 5.5
Anthropic's production-standard Sonnet model for coding, agents, and long-context enterprise workloads.
Claude Fable 5
Anthropic's most capable widely released model for next-generation intelligence and long-running agents.
Claude Opus 5
Anthropic's flagship model for deep reasoning, complex agentic coding, and long-horizon enterprise work.
Claude Sonnet 5
Anthropic's new Sonnet model for extended-context coding, agents, and enterprise production workflows.
Claude Opus 4.8
Anthropic's newest Opus-tier model for complex reasoning, long-horizon agentic coding, and high-autonomy enterprise work.