Claude Haiku 4.5

anthropic/claude-haiku-4.5
VisionToolsJSONReasoning
by Anthropic · 2025-10-15

Claude Haiku 4.5 is Anthropic’s fastest and most efficient model, delivering near-frontier intelligence at a fraction of the cost and latency of larger Claude models. Matching Claude Sonnet 4’s performance...

ctx200K tokens
Max output64K
Inputtext + image + file
Outputtext
p50 TTFT6.58 s
INPUT$1.00/ 1M tokens
OUTPUT$5.00/ 1M tokens
p50 TTFT6.58 s7d
p95 TTFT10.00 s7d
TRAFFIC3.7Mtokens / 7d

Claude Haiku 4.5 is a member of Anthropic's Claude family, specifically optimized for speed and cost. It offers a 200,000-token context window and can generate up to 64,000 output tokens in a single…

What is Claude Haiku 4.5 and who is it for?

How does Claude Haiku 4.5 compare to other Claude models on OrcaRouter?

What modalities can Claude Haiku 4.5 process?

Code samples

Call from any SDK

OpenAI-compatible — keep the SDK you already use

  • OpenAI SDKhttps://api.orcarouter.ai/v1
  • Anthropic SDKhttps://api.orcarouter.ai
from openai import OpenAI

client = OpenAI(
    base_url="https://api.orcarouter.ai/v1",
    api_key="$ORCAROUTER_API_KEY",
)

response = client.chat.completions.create(
    model="anthropic/claude-haiku-4.5",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

Supported parameters

  • include_reasoning
  • max_tokens
  • reasoning
  • response_format
  • stop
  • stream
  • structured_outputs
  • temperature
  • thinking
  • tool_choice
  • tools
  • top_k
  • top_p

Pricing

Input / 1M tokens$1.00
Output / 1M tokens$5.00
Cache read / 1M$0.100
Cache write / 1M$1.25
CurrencyUSD

Cost calculator

Tokens / month10MM
Input share70%%
Estimated / month $22.00 · With prompt caching $18.85

Estimate based on list price

Token & cost estimator

Input tokens: 20Cost per request: $0.002520

Estimate only — actual token counts depend on the provider's tokenizer.

Performance

p50 TTFT
6.58 s
Output speed
127 tok/s
p95 TTFT
10.00 s
Error rate
0%

Public benchmarks

29.6
AA Coding
Better than 31% of models compared
#73 of 106
31.1
AA Intelligence
Better than 31% of models compared
#76 of 110
39.0
AA Math
Better than 17% of models compared
#67 of 81
AIME 2025
39.0
GPQA Diamond
64.6
Humanity's Last Exam
4.3
IFBench
42.0
LiveCodeBench
51.1
Long-Context Recall
43.7
MMLU-Pro
80.0
SciCode
34.4
TerminalBench Hard
27.3
τ²-Bench
32.5
Source: artificialanalysis.ai

How it compares

Claude Haiku 4.5Claude Fable 5Claude Opus 4.7Claude Opus 4.8
Input $/M$1.00$10.00$5.00$5.00
Output $/M$5.00$50.00$25.00$25.00
Context200K1.0M1.0M1.0M
Quality5/1010/1010/1010/10
Compare side-by-sideCompare side-by-sideCompare side-by-sideCompare side-by-side

FAQ

What is the price of Claude Haiku 4.5 on OrcaRouter?
OrcaRouter passes through Anthropic's provider rates with zero markup: $1.00 per 1 million input tokens and $5.00 per 1 million output tokens.
What is the context window size for Claude Haiku 4.5?
Claude Haiku 4.5 has a context window of 200,000 tokens and can output up to 64,000 tokens.
What are Claude Haiku 4.5's main strengths compared to other models?
Its main strengths are speed, low cost, and multimodal input (text, image, file). It has an MMLU-Pro score of 80.0, making it solid for general knowledge tasks while being 3x cheaper than Claude Sonnet on input tokens.
How does Claude Haiku 4.5 compare to GPT-4o Mini?
GPT-4o Mini has a smaller context window (128K vs 200K) and lower max output (16K vs 64K tokens). GPT-4o Mini scores slightly higher on MMLU (~82 vs 80) and is often cheaper. Both are fast lightweight models; choose based on context/output needs and multimodal support.
Does OrcaRouter keep any data when I use Claude Haiku 4.5?
OrcaRouter does not log or store your prompts or generated completions beyond what is necessary to provide the service (e.g., usage metrics for billing). Your data is processed by Anthropic's API under their terms. Refer to OrcaRouter's privacy policy and Anthropic's data handling documentation for details.
How do I call Claude Haiku 4.5 through an OpenAI-compatible API?
Use the base URL https://api.orcarouter.ai/v1, set the model parameter to "anthropic/claude-haiku-4.5", and include your OrcaRouter API key in the Authorization header. The request format follows OpenAI's chat completions endpoint. Example: {"model": "anthropic/claude-haiku-4.5", "messages": [{"role": "user", "content": "Hello"}]}.
Can I use Claude Haiku 4.5 for image analysis?
Yes, Claude Haiku 4.5 accepts image inputs as part of the multimodal support. Provide images in the request as base64 or URL in the content array. The model can describe, classify, or answer questions about images.
What is the token limit for system prompts and long documents?
The total context window is 200,000 tokens, which includes system prompt, user messages, images, files, and any prior conversation history. Each image consumes tokens based on its size (typically ~1,000 tokens for a medium-resolution image). Plan your prompt length accordingly to stay within budget.
Is Claude Haiku 4.5 good for coding tasks?
It can handle common coding patterns, bug fixing, and simple code generation, but it is not as strong as dedicated code models or larger models like Claude Opus. For complex algorithms or full-stack generation, consider a more capable model. Its 64K output token limit allows for large code blocks.
Does Claude Haiku 4.5 support function calling or tool use?
Claude Haiku 4.5 does not natively support OpenAI-style function calling parameters. However, you can prompt it to output structured JSON for simple tool invocation. For robust tool use, use Anthropic's Claude models with tool support (e.g., Claude Sonnet/Opus) or GPT-4o.

Embed this badge

Anthropic: Claude Haiku 4.5$1.00/M in6578ms p50via OrcaRouter
HTML <a href="https://www.orcarouter.ai/models/anthropic/claude-haiku-4.5" target="_blank"> <img src="https://www.orcarouter.ai/embed/anthropic/claude-haiku-4.5.svg" alt="Anthropic: Claude Haiku 4.5 on OrcaRouter" /> </a>
Markdown [![Anthropic: Claude Haiku 4.5](https://www.orcarouter.ai/embed/anthropic/claude-haiku-4.5.svg)](https://www.orcarouter.ai/models/anthropic/claude-haiku-4.5)

Model card as data

GET /api/public/models/anthropic/claude-haiku-4.5Open
Machine-readable:/llms.txt/llms-full.txt