Claude Haiku 5.5
claude-haiku-5-5Compare claude-haiku-5-5 API pricing, supported endpoints, capabilities and access options on Modelsell.
- Starting price
- Input / Output · 1M
- Context
- 1M
- Maximum input window
- Max output
- 128K
- Maximum tokens per response
- Modalities
- →
- Knowledge cutoff
- Jun 2026
- Released
- Oct 2026
Pricing by Supplier
Capabilities / Supported modalities
Provider & data privacy
Anthropic Claude tokenizerPerformance
Benchmarks
Scores on standardized evaluations. Higher percentages are better — and rank percentile shows
Metrics sourced fromArtificial Analysis 2026-09-30·Claude Haiku 5.5
This model is not included in the current benchmark snapshot.
Missing models or measurements are not zero scores.
Benchmark charts preserve the source model selection and reasoning settings. Missing models or measurements are not zero scores, and benchmark cost or speed is not this site’s service commitment.
About Claude Haiku 5.5
-
核心技术规格 • 上下文窗口（Context Window）：1,000,000 tokens（1M） • 最大输出长度（Max Output）：128,000 tokens • 模态支持：支持文本、图片输入，输出为文本 • 知识截止日期：2026 年 6 月 • 自适应思考（Adaptive Thinking）：首次引入 5 档努力程度调节（low、medium、high、xhigh、max，默认 medium），可按需在速度、成本与质量间权衡 • 分词器（Tokenizer）：采用新版分词器（与 Claude 4.7 及之后模型一致），同等文本的 token 数量比 Haiku 4.5 增加约 30%
-
性能大幅跃升 相比上一代 Claude Haiku 4.5，Haiku 5.5 在各项基准测试中均有跨代式提升： • 电脑操作（OSWorld 2.1）：从 15.7% 提升至 72.4%（大幅超越 GPT-6 Luna 的 48.9%） • Agent 编程（Terminal-Bench 4.0）：从 0.0% 提升至 39.2% • 多学科推理（Humanities' Last Exam, 带工具）：从 18.7% 提升至 57.4% • 知识工作（GDPval-AA v2.1）：从 735 分大幅跃升至 1620 分
Use cases and prompting
Starting points for evaluation; supported inputs and options are listed in API access.
Use cases to explore
- Draft and revise text with explicit audience, tone and format requirements.
- Summarize supplied documents and compare answers against the original sources.
Practical tips
Separate instructions from source material, describe the desired output and provide an example. Test tool use or structured output only when listed as supported.
Example prompt
Summarize the following document in five bullet points. Separate confirmed facts from open questions, cite the relevant passages and do not invent missing information. Document: [paste your text]
API access
API documentation
Code samples
/v1/messages| Parameter | Type | Default / range | Description |
|---|---|---|---|
temperature | number | = 10 ~ 2 | Sampling temperature; lower is more deterministic |
top_p | number | = 10 ~ 1 | Nucleus sampling probability mass |
max_tokens | integer | >= 1 | Maximum number of tokens in the response |
frequency_penalty | number | = 0-2 ~ 2 | Penalises repetition of frequent tokens |
presence_penalty | number | = 0-2 ~ 2 | Encourages introducing new topics |
stop | array | — | Up to 4 strings that stop generation |
seed | integer | — | Deterministic sampling seed (best-effort) |
n | integer | = 1>= 1 | Number of completions to generate |
stream | boolean | = false | Stream tokens via Server-Sent Events |
response_format | object | — | Force JSON object or schema-conforming output |
tools | array | — | Tool / function declarations the model may call |
tool_choice | string | autononerequired | Tool-choice policy or specific tool name |
logprobs | boolean | = false | Return per-token log probabilities |
top_logprobs | integer | 0 ~ 20 | Number of top log probabilities returned per token |
logit_bias | object | — | Per-token logit bias map |
user | string | — | End-user identifier for abuse monitoring |
Replace <YOUR_API_KEY> with the API key from your token settings.
Authentication
All requests must include Authorization: Bearer <TOKEN> header. Anthropic-formatted endpoints accept the x-api-key header instead.
Generate tokens from the Tokens page; you can scope them to specific models, groups, IPs, and rate-limits.
Supported parameters
| Parameter | Type | Default / range | Description |
|---|---|---|---|
temperature | number | = 10 ~ 2 | Sampling temperature; lower is more deterministic |
top_p | number | = 10 ~ 1 | Nucleus sampling probability mass |
max_tokens | integer | >= 1 | Maximum number of tokens in the response |
frequency_penalty | number | = 0-2 ~ 2 | Penalises repetition of frequent tokens |
presence_penalty | number | = 0-2 ~ 2 | Encourages introducing new topics |
stop | array | — | Up to 4 strings that stop generation |
seed | integer | — | Deterministic sampling seed (best-effort) |
n | integer | = 1>= 1 | Number of completions to generate |
stream | boolean | = false | Stream tokens via Server-Sent Events |
response_format | object | — | Force JSON object or schema-conforming output |
tools | array | — | Tool / function declarations the model may call |
tool_choice | string | autononerequired | Tool-choice policy or specific tool name |
logprobs | boolean | = false | Return per-token log probabilities |
top_logprobs | integer | 0 ~ 20 | Number of top log probabilities returned per token |
logit_bias | object | — | Per-token logit bias map |
user | string | — | End-user identifier for abuse monitoring |
Rate limits
| Supplier | RPM | TPM | RPD |
|---|---|---|---|
| CCMax | Unlimited | Unlimited | Unlimited |
| CCMax-ZL | Unlimited | Unlimited | Unlimited |
No restriction
Frequently asked questions about claude-haiku-5-5
What is claude-haiku-5-5?
Compare claude-haiku-5-5 API pricing, supported endpoints, capabilities and access options on Modelsell.
How do I call claude-haiku-5-5?
Create an API key with access to claude-haiku-5-5, then use the exact model ID and a supported endpoint from the API access section. Request fields depend on the selected endpoint.
How is claude-haiku-5-5 priced?
Pricing depends on the selected provider group and the model billing unit. The current input, output, request, or media prices are shown on this page before sign-up.
What is the context window of claude-haiku-5-5?
The model catalog lists a context window of 1000000 tokens. Check the selected endpoint for request limits.
What is the maximum output of claude-haiku-5-5?
The model catalog lists a maximum output of 128000 tokens. Your request settings may set a lower limit.
How should I evaluate claude-haiku-5-5 for my project?
Separate instructions from source material, describe the desired output and provide an example. Test tool use or structured output only when listed as supported.