Anthropic: Claude Haiku 4.5

by Anthropic

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 across reasoning, coding, and computer-use tasks, Haiku 4.5 brings frontier-level capability to real-time and high-volume applications. It introduces extended thinking to the Haiku line; enabling controllable reasoning depth, summarized or interleaved thought output, and tool-assisted workflows with full support for coding, bash, web search, and computer-use tools. Scoring >73% on SWE-bench Verified, Haiku 4.5 ranks among the world’s best coding models while maintaining exceptional responsiveness for sub-agents, parallelized execution, and scaled deployment.

Avg Score

78.8%

273 answers

Avg Latency

2.8s

36 runs

Pricing

$1.00

input

$5.00

output

per 1M tokens

Context

200K

tokens

Alternatives

Models with similar or better quality but different tradeoffs

Same Quality, Cheaper

Models with similar or better performance at a lower cost per token.

Model	Cost
Google: Gemini 2.0 Flash	-96%
DeepSeek: DeepSeek V3.2	-92%
Meta: Llama 4 Maverick	-91%
DeepSeek: DeepSeek V3 0324	-87%
Anthropic: Claude 3 Haiku	-87%

Same Quality, Faster

Models with similar or better performance but lower latency.

Model	Latency
Google: Gemini 2.0 Flash	-54%
Google: Gemini 2.5 Flash	-46%
Anthropic: Claude 3 Haiku	-38%
Google: Gemini 3 Flash Preview	-30%
Meta: Llama 4 Maverick	-18%

Same Cost, Better

Models at a similar price point with higher benchmark scores.

Model	Score
Google: Gemini 3 Flash Preview	+11%
MoonshotAI: Kimi K2 Thinking	+2%
DeepSeek: DeepSeek V3.2 Speciale	+1%
DeepSeek: DeepSeek V3 0324	+0%
DeepSeek: DeepSeek V3.2	+0%

Other Models from Anthropic

Compare performance with other models from the same creator

Model	Score	Latency	Cost/1M
Anthropic: Claude Opus 4.5	85.4%	4.0s	$15.00
Anthropic: Claude Sonnet 4.5	82.7%	3.9s	$9.00
Anthropic: Claude 3.5 Haiku	77.7%	3.3s	$2.40
Anthropic: Claude 3 Haiku	75.2%	1.8s	$0.75
Anthropic: Claude Opus 4.1	—	—	$45.00
Anthropic: Claude Opus 4	—	—	$45.00
Anthropic: Claude Sonnet 4	—	—	$9.00
Anthropic: Claude 3.7 Sonnet	—	—	$9.00
Anthropic: Claude 3.7 Sonnet	—	—	$9.00
Anthropic: Claude 3.5 Haiku (2024-10-22)	—	—	$2.40
Anthropic: Claude 3.5 Sonnet	—	—	$18.00

Benchmark Performance

How this model performs across different benchmarks

Benchmark	Score	Rank
moose	100.0%	5 / 17
Math Bench: Subtraction (1-10)	100.0%	31 / 35
Enhancing FIDO UV Consistency with UXWG 2026 priority plan	100.0%	11 / 23
Math Bench: Arithmetic (1-100)	100.0%	7 / 35
Customer service	100.0%	14 / 17
Math Bench: Division (1-10)	100.0%	14 / 35
Testing_website	100.0%	15 / 17
solve n-queens in clojure in one shot	100.0%	2 / 17
Best bathroom technique	100.0%	14 / 17
Breakfast	100.0%	7 / 17
Repetitive Onomatopoeia Response Benchmark	100.0%	14 / 17
Math Bench: Addition (1-10)	100.0%	31 / 35
Cheeky_test	100.0%	14 / 17
Motor production test	100.0%	18 / 36
Customer Support Ticket Generation	100.0%	14 / 17
Math Bench: Multiplication (1-10)	100.0%	31 / 35
Structured Customer Support Extraction	100.0%	14 / 17
Smell test	100.0%	15 / 17
SaaS Customer Support Extraction and Strategy	100.0%	3 / 17
Venture Capital Terms Benchmark	99.7%	13 / 25
Spatial Reasoning: Germany	96.5%	2 / 35
Define data taxonomy for product catalog suitable for PIM	95.0%	8 / 17
KUKA force Torque control programming skills	95.0%	6 / 17
Character Frequency Bench	94.5%	5 / 35
Customer support	90.0%	10 / 17
Math Bench: Arithmetic (1-1000)	90.0%	26 / 35
Product Recommendations	83.3%	7 / 17
Karlsruhe Local Knowledge Benchmark	66.1%	26 / 35
Product Recommendations	61.9%	27 / 36
Rust performance	50.0%	5 / 16
Current Global Political Leaders	30.0%	2 / 17
German Memelord Bench	15.9%	19 / 35
Money Boy Cultural Literacy Test	0.0%	33 / 35
Karl Lorey knowlege	0.0%	31 / 36

Price vs Performance

Compare cost efficiency across all models

Current model

Other models

X-axis uses log scale for better visualization of price range

Score Over Time

Performance trends across all benchmark runs

Benchmark Activity

Number of benchmark runs over time

Quickstart

Get started with this model using OpenRouter

View on OpenRouter

import { OpenRouter } from "@openrouter/sdk";

const openrouter = new OpenRouter({
  apiKey: "<OPENROUTER_API_KEY>"
});

const completion = await openrouter.chat.completions.create({
  model: "anthropic/claude-haiku-4.5",
  messages: [
    {
      role: "user",
      content: "Hello!"
    }
  ]
});

console.log(completion.choices[0].message.content);

Get your API key at openrouter.ai/keys