ZMIME
Anthropic · Released Oct 22, 2024

Claude Haiku 3.5

Fast Claude model for responsive assistance, classification, and lightweight agents.

  • Proprietary
  • Tool calling
  • Vision
Capability
124.8ECI · #145 of 158
Input
$0.80per 1M tokens
Output
$4.00per 1M tokens
Context
200K8K max output
Pricing

What it costs

Input
$0.80per million tokens
Output
$4.00per million tokens
Blended (3:1)
$1.60pricier than 68% of priced models

Typical monthly bills

Monthly usageEstimated cost
Side project2M input + 0.5M output tokens$3.60
Team assistant25M input + 5M output tokens$40.00
Production app250M input + 50M output tokens$400.00

Median of 3 paid API providers on models.dev.

Capability

Independent benchmarks

124.8Epoch Capabilities Index
#145 of 158 scored models
68Median 147174

The shaded band is Epoch AI’s confidence range (121.5–127.7); the tick marks the median scored model.

  • GPQA DiamondGraduate-level science questions38.1%
  • OTIS Mock AIME 2024–2025Competition mathematics4.3%

Scores from Epoch AI, run independently of Anthropic.

Specifications

The details

Lab
Anthropic
Released
Oct 22, 2024
Knowledge cutoff
Jul 31, 2024
Context window
200,000 tokens
Max output
8,192 tokens
Inputs
Text, Images, PDFs
Output
Text
Reasoning
No
Tool calling
Yes
Structured output
No
Weights
Proprietary
Availability
4 API providerslisted on models.dev
Lab-reported

What Anthropic claims

Published by the lab at launch. Settings vary, so compare these only with care.

BenchmarkScoreSettingSource
Aider Polyglot28 percent correct—Source
Same lab

More from Anthropic

Questions

About Claude Haiku 3.5

How much does Claude Haiku 3.5 cost?

Claude Haiku 3.5 costs $0.80 input / $4.00 output per million tokens (median across 3 API providers). At a 3:1 input-to-output mix that is $1.60 per million tokens, more expensive than 68% of the 469 priced models we track.

What is the context window of Claude Haiku 3.5?

Claude Haiku 3.5 accepts up to 200,000 tokens per request and can write up to 8,192 tokens in one response.

How good is Claude Haiku 3.5?

Epoch AI gives Claude Haiku 3.5 a Capabilities Index score of 124.8 (likely range 121.5–127.7), ranking it #145 of 158 models Epoch has scored. Epoch AI benchmark results: GPQA Diamond 38.1%, OTIS Mock AIME 2024–2025 4.3%.

Is Claude Haiku 3.5 open source?

No. Claude Haiku 3.5 is proprietary; you use it through an API or partner platforms.

What inputs does Claude Haiku 3.5 support?

Claude Haiku 3.5 accepts text, images and PDFs and replies in text. It is not a dedicated reasoning model, supports tool calling.

When was Claude Haiku 3.5 released?

Anthropic released Claude Haiku 3.5 on Oct 22, 2024. Its training data runs to Jul 31, 2024.

What are the best alternatives to Claude Haiku 3.5?

The closest current models from other labs on capability, price and release date are Llama-3.1-70B-Instruct (Meta, ECI 124.7, $0.72 / $0.72), Mistral Large 2407 (Mistral AI, ECI 126.4, $2.00 / $6.00), GPT-4o (OpenAI, ECI 128.1, $2.50 / $10.00) and Gemma 2 27B (Google, ECI 118.6, $0.65 / $0.65).