anthropic/claude-haiku-4-5

Aliases: haiku-4.5, haiku

$1.00

Input / 1M

$5.00

Output / 1M

200K

Context

165 t/s

Speed

Public benchmark scores

Sourced from each provider's published numbers. Verify before quoting.

Quality index

MMLU

81.2

GPQA

50.1

HumanEval

78.3

MATH

76.5

SWE-bench

—

Arena Elo

—

Sources: anthropic-system-card

Capabilities

toolsvisionstreamingstructured_outputprompt_cache

Input modalities: text, image

Use claude-haiku-4-5 via Relay

Configure the model alias in YAML, then call it from Python.

YAML

# models.yaml
version: 1
models:
  haiku-4.5:
    target: anthropic/claude-haiku-4-5
    credential: $env.ANTHROPIC_API_KEY

Python

from relay import Hub

async with Hub.from_yaml("models.yaml") as hub:
    resp = await hub.chat(
        "haiku-4.5",
        messages=[{"role": "user", "content": "Hello"}],
    )
    print(resp.text, resp.cost_usd)

pip install ai5labs-relay · full docs on GitHub

Compare with

vs mistral/mistral-large-latest vs openai/gpt-4o-mini vs groq/llama-3.3-70b-versatile vs anthropic/claude-3-5-sonnet-20241022