All models

Claude Haiku 4.5

Anthropic's fast, cheap model. The right choice for high-volume tasks where latency and cost dominate.

Pricing

Input
$1 / 1M
Output
$5 / 1M

Billed per token. No subscription, no minimums.

Specs

Context window
200,000 tokens
Max output
8,192 tokens

Capabilities

chatvisiontool-use

Modalities

textimage

About Claude Haiku 4.5

Claude Haiku 4.5 is the smallest and fastest model in the Claude family. It's the right pick for high-volume operations — classification, light extraction, routing, first-pass summarisation, embedding-adjacent text tasks — where each call needs to be cheap and fast.

Haiku shares the Claude family's tool-calling format and vision support, so it can act as the cheap front-end in a tiered routing setup: Haiku triages or pre-processes, and only the requests that need it get handed up to Sonnet or Opus. The 200K context window holds for Haiku too, which means you don't have to design around short context just because you're optimising for cost.

Try it

Call claude-haiku-4-5-20251001 through Troviale with your existing OpenAI SDK — just swap the base URL and key.

client = OpenAI(
  base_url="https://api.troviale.com/v1",
  api_key="YOUR_TROVIALE_KEY",
)

client.chat.completions.create(
  model="claude-haiku-4-5-20251001",
  messages=[{"role": "user", "content": "Hello"}],
)

Pricing: input $1 / 1M tokens · output $5 / 1M tokens. Prices shown as published — actual billing reflects token counts returned by the provider.