Anthropic has announced that its new and improved Claude 3.5 Haiku AI model is going up in price – only two weeks after it was released into the world.
The most affordable option in Anthropic’s range of three models was originally expected to retain the same pricing of the previous 3 Haiku product. But that cost will now increase fourfold, at $1 per million input tokens and $5 per million output tokens.
Claude 3.5 Haiku offers iterative updates and improvements on its predecessor. This “increase in intelligence”, Anthropic says, is so significant as to justify the inflation in price.
Haiku Price Hiked
California-based artificial intelligence company Anthropic announced on 22nd October that Claude 3.5 Haiku would replace the previous version 3, along with news of upgrades to its Claude 3.5 Sonnet model and public beta release of its ‘computer use’ tool.
But the announcement was subsequently updated on 4th November to confirm that: “We have revised the pricing for Claude 3.5 Haiku. The model is now priced at $1 MTok input / $5 MTok output.”
This just in! View
the top business tech deals for 2024 👨💻
The reason for the Haiku price hike was briefly explained further in a post from the company on X, formerly Twitter.
During final testing, Haiku surpassed Claude 3 Opus, our previous flagship model, on many benchmarks—at a fraction of the cost.
As a result, we’ve increased pricing for Claude 3.5 Haiku to reflect its increase in intelligence: https://t.co/UWTd1RDRVE.
— Anthropic (@AnthropicAI) November 4, 2024
Anthropic calls Haiku its “fastest, most cost-effective model”, with its Sonnet and Opus models costing three and 15 times more expensive respectively.
Anthropic’s Pricing Explained
Instead of working on a flat subscription rate or charging by interaction, Anthropic uses a token basis for working out the users’ bill – this method is common to several high profile AI companies, including OpenAI’s pricing for ChatGPT.
While Claude 3 Haiku cost $0.25 per million input tokens and $1.25 per million output tokens, that has gone up to $1 per million and $5 per million respectively for its successor.
Words, images, lines of code, PDF files and other forms of input and output are all attributed a token value. The cost is then worked out by how many tokens are used for the user’s inputs (i.e. prompts, questions and commands) and outputs (the results).
One of the best AI productivity tools on the market, the pricing structure reflects the fact that Haiku is aimed more towards businesses and teams of collaborators than it is individual users.
What’s New for Haiku?
Anthropic’s announcement for the new and improved Haiku model was light on impressive sounding new features, but robust in its claims of increased power:
“For a similar speed to Claude 3 Haiku, Claude 3.5 Haiku improves across every skill set and surpasses even Claude 3 Opus, the largest model in our previous generation, on many intelligence benchmarks.” – Anthropic
The product launch announcement says that it is particularly strong for coding tasks, with benchmarking that sees it outperform Open AI’s GPT-4o.
“With low latency, improved instruction following, and more accurate tool use,” says the statement, “Claude 3.5 Haiku is well suited for user-facing products, specialized sub-agent tasks, and generating personalized experiences from huge volumes of data—like purchase history, pricing, or inventory records.”