Claude Haiku 4.5 API

anthropic/claude-haiku-4.5
200K tokens · 1000 input / 5000 output credits / 1M tokens

Claude Haiku 4.5 Chat turns text and image messages into concise answers, code suggestions, and extracted information for repeated tasks. Its 200K-token context allows relevant instructions and source material to travel together, making it useful for responsive workflows.

Claude Haiku 4.5

Anthropic · chat-completions

Chat with Claude Haiku 4.5

Each model has its own conversation. Switching never sends another model's history, and switching back resumes where you left off. Requests are billed from actual token usage.

Ctrl / ⌘ + Enter to send0 / 32,000

Continue with

Claude Haiku 4.5

Claude Haiku 4.5 handles frequent conversational tasks such as short coding assistance, extraction, and image questions. Send ordered text or image messages with the relevant instructions, then use max_tokens to set the response length. Its 200K-token context supports substantial source material, while the Claude Messages response includes token usage for each request. Choose /v1/chat/completions for an ordered messages array with a system role, or /v1/messages for a separate system field. Both endpoints return a complete JSON response by default.

Why Choose Claude Haiku 4.5?

  • Responsive task cyclesGet concise answers for repeated questions, classification, and extraction workflows.

  • Practical coding assistanceAsk for focused code suggestions, explanations, or small changes with the relevant snippet attached.

  • Text and image inputUse written context and image blocks together when a task depends on visual material.

Parameters

/v1/chat/completions

ParameterRequirementDescription
modelRequired

Use anthropic/claude-haiku-4.5.

messagesRequired

A non-empty ordered array of conversation messages; each item contains role and content.

max_tokensOptional

A positive integer setting the response output limit; the Playground offers 1?65,536.

Default4096
temperatureOptional

Sampling temperature; the Playground offers 0?2.

Default1
top_pOptional

Nucleus sampling probability; the Playground offers 0?1.

Default1
streamOptional

Set true for an SSE event stream; the Playground uses complete JSON responses.

Defaultfalse

/v1/messages

ParameterRequirementDescription
modelRequired

Use anthropic/claude-haiku-4.5.

messagesRequired

A non-empty ordered array of conversation messages; each item contains role and content.

systemOptional

System instructions supplied separately from messages.

max_tokensRequired

A positive integer setting the response output limit; the Playground offers 1?65,536.

Default4096
temperatureOptional

Sampling temperature; the Playground offers 0?2.

Default1
top_pOptional

Nucleus sampling probability; the Playground offers 0?1.

Default1
streamOptional

Set true for an SSE event stream; the Playground uses complete JSON responses.

Defaultfalse

How to Use

  1. Choose the endpointSelect /v1/chat/completions or /v1/messages and set model to anthropic/claude-haiku-4.5.

  2. Provide the conversationAdd ordered messages. Put system guidance in a system-role message for Chat Completions, or in the system field for Messages.

  3. Set the response lengthSet max_tokens to a positive integer for Messages, then send the request and review the response and token usage.

Pricing

Vidgo AI measures input and output tokens separately. Each rate applies per 1M tokens.

UsageRateDetails
Input$0.80 · 1000 credits / 1M tokensTokens in the request.
Output$4.00 · 5000 credits / 1M tokensTokens generated in the response.

Best Use Cases

  • High-volume extractionExtract named facts or short summaries from repeated inputs.

  • Focused code helpExplain a function, suggest a local change, or draft a small test.

  • Image questionsDescribe the relevant details in a chart, screenshot, or scene.

Pro Tips

  • Ask for the exact fields or answer format needed by your application.
  • Include the source snippet or image that the answer must rely on.
  • Set max_tokens to a value that fits the expected response length.

Usage Notes

  • Claude Haiku 4.5 has a 200K-token context window and a 64K-token model output limit.
  • The Playground max_tokens control ranges from 1 to 65,536 and starts at 4,096.
  • Use anthropic/claude-haiku-4.5 with /v1/chat/completions or /v1/messages.

Related Models

Claude Haiku 4.5 API frequently asked questions

What is the Claude Haiku 4.5 API?

Claude Haiku 4.5 turns conversation text and image context into useful answers. Its 200K-token context supports connected work. Use anthropic/claude-haiku-4.5 through Chat Completions or Claude Messages, or try the Playground above.

What tasks fit Claude Haiku 4.5?

Claude Haiku 4.5 fits repeated extraction, short explanations, focused code suggestions, and image questions. Supply the specific material and requested result for each task.

What context and output sizes does Claude Haiku 4.5 have?

Claude Haiku 4.5 has a 200K-token context window and a 64K-token model output limit. In the Playground, max_tokens ranges from 1 to 65,536.

How does Claude Haiku 4.5 read images?

Put an image block in messages.content and add a text question about it. The model returns its answer as text.

What does Claude Haiku 4.5 cost on Vidgo AI?

Claude Haiku 4.5 is $0.80 and 1000 credits per 1M input tokens, and $4.00 and 5000 credits per 1M output tokens. The Pricing section shows both rates.

Which API endpoints can I use for Claude Haiku 4.5?

Use anthropic/claude-haiku-4.5 with /v1/chat/completions or /v1/messages. Choose Chat Completions for system-role messages or Messages for a separate system field.