Topics
in style
AI
Amazon
Image Credits:Anthropic
Apps
Biotech & Health
Climate
Cloud Computing
Commerce
Crypto
Enterprise
EVs
Fintech
fundraise
gizmo
Gaming
Government & Policy
Hardware
layoff
Media & Entertainment
Meta
Microsoft
Privacy
Robotics
Security
societal
distance
inauguration
TikTok
Transportation
speculation
More from TechCrunch
Events
Startup Battlefield
StrictlyVC
Podcasts
Videos
Partner Content
TechCrunch Brand Studio
Crunchboard
Contact Us
Anthropic , one of the world ’s big AI vendor , has a powerful family of generative AI models predict Claude . These models can perform a range of project , from captioning images and write electronic mail to solving maths and coding challenges .
With Anthropic ’s model ecosystem growing so quickly , it can be tough to keep track of which Claude poser do what . To aid , we ’ve put together a guidebook to Claude , which we ’ll keep updated as new model and raise arrive .
Claude models
Claude manikin are named after literary works of art : Haiku , Sonnet , and Opus . The latest are :
Counterintuitively , Claude 3 Opus — the big and most expensive example Anthropic go — is the least capable Claude model at the second . However , that ’s sure to change when Anthropic publish an update variant of Opus .
Most recently , Anthropic releasedClaude 3.7 Sonnet , its most advanced model to date . This AI good example is different from Claude 3.5 Haiku and Claude 3 Opus because it ’s a intercrossed AI abstract thought model , which can give both material - sentence solution and more weigh , “ thought - out ” answers to questions .
When using Claude 3.7 Sonnet , users can choose whether to sprain on the AI model ’s reasoning abilities , which prompt the model to “ opine ” for a short or long menstruum of time .
When logical thinking is plough on , Claude 3.7 Sonnet will spend anywhere from a few seconds to a couple second in a “ thinking ” phase before answering . During this phase , the AI model is breaking down the substance abuser ’s prompting into small parts and checking its answers .
Join us at TechCrunch Sessions: AI
Exhibit at TechCrunch Sessions: AI
Claude 3.7 Sonnet is Anthropic ’s first AI model that can “ argue , ” a techniquemany AI labs have turned to as traditional methods of improving AI performance wick off .
Even with its logical thinking incapacitate , Claude 3.7 Sonnet remains one of the tech industry ’s top - do AI manakin .
In November , Anthropic released an improved — and more expensive — version of its lightweight AI framework , Claude 3.5 Haiku . This model outstrip Anthropic ’s Claude 3 Opus on several benchmark , but it ca n’t break down images like Claude 3 Opus or Claude 3.7 Sonnet can .
All Claude poser — which have a standard 200,000 - token context window — can also follow multistep didactics , use tools(e.g . , stock ticker trackers ) , and produce integrated output in formats likeJSON .
A context window is the amount of data a model like Claude can dissect before yield unexampled data , while tokens are subdivide act of raw data ( like the syllable “ devotee , ” “ tas , ” and “ tic ” in the word “ howling ” ) . Two hundred thousand tokens is tantamount to about 150,000 words , or a 600 - page novel .
Unlike many major productive AI model , Anthropic ’s ca n’t get at the internet , meaning they ’re not particularly great at answering current events questions . They also ca n’t generate images — only wide-eyed line diagram .
As for the major difference between Claude exemplar , Claude 3.7 Sonnet is firm than Claude 3 Opus and better realize nuanced and complex statement . Haiku struggles with advanced prompts , but it ’s the swiftest of the three mannikin .
Claude model pricing
The Claude models are usable through Anthropic ’s API and managed platform such asAmazon Bedrockand Google Cloud’sVertex AI .
Here ’s the Anthropic API pricing :
Anthropic offers quick hive up and batching to move over extra runtime delivery .
Prompt caching lease developers store specific “ straightaway context ” that can be reused across API calls to a good example , while batching process asynchronous groups of low - anteriority ( and subsequently cheaper ) model illation requests .
Claude plans and apps
For individual users and companies looking to but interact with the Claude models via apps for the web , Android , and iOS , Anthropic offer a free Claude plan with rate limit point and other usance restrictions .
promote to one of the party ’s subscriptions take away those limits and unlocks new functionality . The current plans are :
Claude Pro , which costs $ 20 per month , make out with 5x eminent charge per unit limits , antecedency admission , and previews of upcoming feature .
Being business - concentrate , Team — which be $ 30 per drug user per month — add a dashboard to check billing and user management and integrations with information repos such as codebases and client relationship management platforms ( for instance , Salesforce ) . A toggle enable or disables citations to verify AI - generate claims . ( Like all manakin , Claudehallucinatesfrom sentence to time . )
For customers who need even more , there ’s Claude Enterprise , which allows companies to upload proprietary data into Claude so that Claude can analyze the info and suffice questions about it . Claude Enterprise also comes with a heavy context window ( 500,000 tokens ) , GitHub integration for engineering teams to sync their GitHub repository with Claude , and Projects and Artifacts .
A word of caution
As is the case with all productive AI models , there are risks relate with using Claude .
The manikin occasionallymake mistakes when summarizingoranswering questionsbecause of their tendency tohallucinate . They ’re also trained on public vane data , some of which may be copyrighted or under a restrictive license . anthropical and many other AI vendors contend that thefair - usedoctrine shield them from copyright claim . But that has n’t stopped information ownersfromfiling lawsuits .
Anthropicoffers policiesto protect certain customers from courtroom battles spring up from honest - use challenge . However , they do n’t correct the ethical quandary of using modelling train on datum without permit .
This article was originally published on October 19 , 2024 . It was update on February 25 , 2025 , to include new detail about Claude 3.7 Sonnet and Claude 3.5 Haiku .
