The fastest human model, Cloud 3.5 Haiku, is now generally available

Anthropic has officially rolled out the Claude 3.5 Haiku model to all users through its Claude chatbot on the web and mobile apps. AI users see it on X.

Previously limited to developers accessing it via Anthropic’s API after its launch in October 2024, this smaller, faster model has received attention for its ability to outperform larger models on key benchmarks while maintaining a competitive price point.

According to a third-party benchmarking organization Synthetic analysisCloud 3.5 Haiku “has lower latency compared to the average, taking 0.80 seconds to receive the first token (TTFT),” however it is “slower compared to the average, with an output speed of 65.1 tokens per second.”

The release – which has not been officially announced – comes on the heels of major updates from Anthropic’s AI competitors OpenAI and Google, who also shipped new models for general availability in their chatbots as the year winds down, namely OpenAI’s o1 and o1-mini models and Google Gemini 2.

The question for Anthropic is whether customers will be impressed enough with the Claude 3.5 Haiku’s performance to sign up for its Pro tier — or continue to use it instead of some of these other advanced, fast competitors.

Claude 3.5 Haiku can be accessed through the Claude Chatbot

As the fastest and most cost-effective model in Anthropic’s lineup, the Claude 3.5 Haiku excels at real-time tasks such as processing large data sets, analyzing financial documents, and generating output from long-form context information.

It features a context window containing 200,000 symbols – more than the 128,000 symbol window in OpenAI’s GPT-4 – allowing it to handle extensive input with ease.

On the Claude chatbot, Haiku provides functionality that enhances its versatility. Users can analyze images and file attachments, making it useful for multimedia tasks and workflows involving large collections of documents.

Haiku also integrates with Claude Antiquesan interactive sidebar that debuted in June 2024. Artifacts provides a dedicated workspace to process and enhance AI-generated content in real time, including running full applications. In my testing of Artifacts with Haiku this morning, I was able to program a fully playable version of Pong in less than a minute:

Despite its strengths, haiku has limitations. It does not currently support web browsing or image creation, both of which are offered by competitors such as OpenAI’s GPT-4o and GPT-4.

Additionally, my brief test of it this morning showed that it failed the “strawberry test,” a popular user-designed challenge in which an AI must identify all three r’s in the word strawberry.

Access and subscription details

Claude 3.5 Haiku is accessible for free via the Claude chatbot, but users face a variable daily message limit depending on server demand.

For example, in the free tier this morning when I tried it, I was able to do about 10 exchanges (20 total incoming and outgoing messages) before hitting Anthropic’s quota, which resets daily.

To unlock more comprehensive usage, users can subscribe to the Claude Pro plan, priced at $20 per month.

This subscription provides up to five times the usage of the free tier, priority access during high traffic periods, early access to new features, and access to additional samples such as Claude 3 Opus.

The pricing structure mirrors OpenAI’s ChatGPT Plus subscription, providing a premium experience for power users.

Performance and cost

On the API, the Claude 3.5 Haiku delivers exceptional performance at an affordable price. Starting at $0.80 per million input symbols and $4 per million output symbols, it provides an economical solution compared to larger models such as the Claude 3 Opus.

Developers can further reduce costs by using Fast Caching, which provides up to 90% savings, and the Message Batches API, which reduces costs by 50%.

In benchmark tests, the Haiku outpaced many larger, publicly available models. Its performance includes a score of 40.6% on the SWE-bench Verified test, a key coding benchmark, demonstrating its strength in tasks that require intelligence and speed. This makes Haiku an excellent choice for user-facing applications and time-sensitive workflows.

Key considerations

While Claude 3.5 Haiku offers powerful capabilities, potential users should consider current limitations. The lack of web browsing and image generation may make it less attractive for certain use cases compared to competitors. Furthermore, the daily message limit may be inconvenient for users who do not want to upgrade to a Claude Pro subscription.

However, with features like image and file analysis, powerful markup capabilities, and integration with Artifacts, Haiku remains a powerful tool for tasks that require speed and accuracy.

The Artifacts feature, in particular, extends its functionality beyond text generation, enabling collaborative editing and content optimization in real time.

For users ready to explore its capabilities, Claude 3.5 Haiku is now live and available through the Claude chatbot on the web and mobile apps on iOS and Android.