Anthropic
Claude Haiku 5.5
anthropic/claude-haiku-5.5
Anthropic's small, fast model for high-volume, cost-sensitive work such as summarization, subagents, and browser use, with stronger coding and computer use than Haiku 4.5.
RoleThe subagent and bulk worker
$0.100
In / M
$0.500
Out / M
1.0M
Context
- AA intelligence index
- 36.5
- Inputs
- text, image, file
- Outputs
- text
- Tools
- Listed
- Structured output
- Listed
- Reasoning
- Listed
- Weights
- Not verified
- Listed
- Oct 7, 2026
01 · What it's good at
Use it for this.
Three jobs this model is built for, and the ones to give to something else.
01
High-volume summarization and extraction
02
Subagents inside a bigger agent
03
Browser-use and computer-use steps
Skip it for
- The hardest reasoning in your pipeline. Escalate to Sonnet or Opus
02 · Example tasks
What you'd actually send it.
Real inputs, not a benchmark. Paste one into the snippet below to try it.
You send
Extract name, company, intent, and budget from this inbound email. Return JSON only.
Why this model
Structured output at volume is the classic small-model job.
03 · Your stack
Where it sits: Do the volume.
Put it wherever you call a model many times: per email, per ticket, per page, or as a subagent spawned by your main agent. It is also the chat model in OpenRouter's push-to-talk voice cookbook.
Layer 01
Retrieve
Layer 02
Plan & reason
Layer 03 · this model
Do the volume
Layer 04
Gate & route
Layer 05
Draw, speak, listen
04 · Wire it up
One request. Paste and run.
Uses OpenRouter, so one key works for every model on this site. Set OPENROUTER_API_KEY and swap the prompt for one of the tasks above.
// OpenRouter speaks the OpenAI Chat Completions format.
const res = await fetch("https://openrouter.ai/api/v1/chat/completions", {
method: "POST",
headers: {
Authorization: `Bearer ${process.env.OPENROUTER_API_KEY}`,
"Content-Type": "application/json",
},
body: JSON.stringify({
"model": "anthropic/claude-haiku-5.5",
"messages": [
{
"role": "user",
"content": "Extract the lead from this email: Hi, I'm Dana from Acme. We need SSO for 200 seats, budget around $40k."
}
],
"reasoning": {
"effort": "none"
},
"response_format": {
"type": "json_schema",
"json_schema": {
"name": "lead",
"strict": true,
"schema": {
"type": "object",
"properties": {
"name": {
"type": "string"
},
"company": {
"type": "string"
},
"intent": {
"type": "string"
},
"budget_usd": {
"type": [
"number",
"null"
]
}
},
"required": [
"name",
"company",
"intent",
"budget_usd"
],
"additionalProperties": false
}
}
}
}),
});
const data = await res.json();
const result = JSON.parse(data.choices[0].message.content);
console.log(result);
05 · Watch out for
Gotchas.
Reasoning is optional here. Effort
noneturns it off, which is the fast path for extraction.The listing has a higher price tier for prompts above 100K tokens. Keep bulk prompts short.
Sources
Lessons are written from each model's listing and the linked docs. We don't publish our own benchmark numbers; the index above is Artificial Analysis, as shown on OpenRouter.
Keep learning