Anthropic
Claude Haiku 5.5
Anthropic's latest Haiku for low-latency, high-volume workloads, with 1M context and pricing tiers at 100K input tokens
Context Length1M tokens
Max Output128K tokens
MultimodalText and image input; text output
Pricing & Specs
๐ฐ Pricing
Input$0.10 / M tokens
Output$0.50 / M tokens
Cache Write$0.20 / M tokens
Cache Hit$0.01 / M tokens
โ๏ธ Specs
Context Length1M tokens
Max Output128K tokens
MultimodalText and image input; text output
Adaptive ThinkingOn by default; omit thinking or use adaptive, and control depth with output_config.effort (default: medium)
Short Context Pricing (input โค100K)Input $0.10; cache read $0.01; cache write $0.20; output $0.50 / MTok
Long Context Pricing (input >100K)Input $0.50; cache read $0.05; cache write $1; output $2.50 / MTok; these rates apply to the entire request
Pricing ThresholdTotal input includes uncached input, cache reads, and cache writes; exactly 100,000 tokens still uses short-context rates
Cache Write BillingThis service bills both 5m and 1h cache writes at the 1h rate: $0.20 / MTok for short context, $1 / MTok for long context
Sampling ParametersOmit temperature, top_p, and top_k; replace thinking.type=enabled and budget_tokens with adaptive thinking
Tool CallingSupports auto, any, and named tools; forced tool calls return without a thinking block
API CompatibilityClaude Messages API; Claude Code
Release DateOctober 7, 2026
Knowledge CutoffJune 2026
API Examples
Python
import anthropic
client = anthropic.Anthropic(api_key="YOUR_API_KEY", base_url="https://api.zairouter.com")
message = client.messages.create(
model="claude-haiku-5-5",
max_tokens=4096,
messages=[{"role": "user", "content": "Hello, Claude!"}],
)
for block in message.content:
if block.type == "text":
print(block.text)
cURL
curl https://api.zairouter.com/v1/messages \
-H "x-api-key: YOUR_API_KEY" \
-H "anthropic-version: 2023-06-01" \
-H "Content-Type: application/json" \
-d '{"model":"claude-haiku-5-5","max_tokens":4096,"messages":[{"role":"user","content":"Hello, Claude!"}]}'
Developer Assist
export ANTHROPIC_AUTH_TOKEN=YOUR_API_KEY
export ANTHROPIC_BASE_URL=https://api.zairouter.com
claude --model claude-haiku-5-5
Official sources (verified October 8, 2026): Claude Haiku 5.5 ยท Anthropic Pricing.
When upgrading from Haiku 4.5, follow the official migration guide. Responses may start with thinking blocks, so the example selects text by type; allow room for both thinking and text in max_tokens.