Artificial intelligence company Anthropic’s latest generative AI model, Claude 3.5 Sonnet, is already performing well against leading rivals like OpenAI’s GPT-4o on key AI benchmarks days after release, the latest advance as companies race to develop bigger and better AI models and capitalize on tech’s latest gold rush.
Claude is Anthropic’s ChatGPT rival.
SOPA Images/LightRocket via Getty Images
Key Takeaways
- Anthropic released Claude 3.5 Sonnet last Thursday, boasting it is “our most intelligent model yet” and capable of outperforming competitors like Meta’s Llama and Google’s Gemini on “key evaluations” like reasoning and coding while running at twice the speed and for one-fifth the cost of its previous flagship model Claude 3 Opus.
- Anthropic pitches Sonnet as the “hard-working” middle child for “efficient, high-throughput tasks” within its three-person family of Claude AI models, combining aspects of its speedy Haiku model and its “powerful” Opus model designed for highly complex tasks.
- The models power Anthropic’s chatbot Claude, a rival to OpenAI’s popular ChatGPT that is available as both a free and paid-for “pro” product.
- Just days after its release, Claude 3.5 Sonnet has already climbed to the top of various league tables of the LMSYS Chatbot Arena, a major crowdsourced benchmarking site for large language models, and it handily outperformed its older Anthropic sibling Claude Opus 3.
- In a post on X, formerly Twitter, the LMSYS account said Claude 3.5 Sonnet had “made a huge leap,” securing pole position for coding and hard prompts, though it ranked second on LMSYS’ overall leaderboard assessing the full range of AI functions behind OpenAI’s newly released model GPT-4o.
- For most people using the AI tool for everyday tasks, the difference in performance is likely to be so slight as to be unnoticeable and the LMSYS account stressed Claude 3.5 Sonnet was up there with the best and “competitive… across the boards” with leading frontier models like GPT-4o and Gemini 1.5 Pro.
What Is Anthropic?
Anthropic is a U.S. AI company and a key player in the generative AI landscape. It pitches itself as a safety-focused firm that prioritizes the safe and responsible development of artificial intelligence over commercialization and profits. Anthropic was founded in 2021 by a group of researchers defecting from industry leader OpenAI over concerns the company cared more about making money than safety.
This cohort included siblings and former OpenAI executives Dario Amodei and Daniela Amodei, who respectively lead the company today as CEO and president. It has secured backing totaling $8.36 billion, including major deals with titans like Amazon and Google, and has a post-money valuation of $18.4 billion after raising a according to Pitchbook data.
What To Watch For
Anthropic said Claude 3.5 Sonnet is “our first release in the forthcoming Claude 3.5 model family.” The AI company said it plans to release Claude 3.5 Haiku and Claude 3.5 Opus “later this year,” though it did not specify a date or rough timeline. Based on the capabilities of the Claude family, it can be expected that Haiku will be the least powerful but fastest of the trio, and Opus most powerful Claude AI model within Anthropic’s stable. The company said its “aim is to substantially improve the tradeoff curve between intelligence, speed, and cost every few months,” adding that it is developing new modalities and features to support businesses and enterprise applications in addition to developing the next-gen models.
This article was first published on forbes.com and all figures are in USD.
Are you – or is someone you know -creating the next Afterpay or Canva? Nominations are open for Forbes Australia’s first 30 under 30 list. Entries close midnight, July 15, 2024.
Look back on the week that was with hand-picked articles from Australia and around the world. Sign up to the Forbes Australia newsletter hereor become a member here.
