Anthropic, OpenAI’s biggest rival, has launched its latest AI model called Claude 3.5 Sonnet — the company’s first release in the upcoming Claude 3.5 AI model series. Anthropic has claimed that its latest offering outperforms its peers such as OpenAI’s GPT-4o, Google’s Gemini-1.5 Pro, Meta’s Llama-400b, and even the company’s proprietary models — Claude 3 Haiku and Claude 3 Opus. “Claude 3.5 Sonnet operates at twice the speed of Claude 3 Opus. This performance boost, combined with cost-effective pricing, makes Claude 3.5 Sonnet ideal for complex tasks such as context-sensitive customer support and orchestrating multi-step workflows,” Anthropic said in a statement. What is Claude 3.5 Sonnet? Claude 3.5 Sonnet is a large language model (LLM), and is part of the family of LLMs which is being developed by Anthropic. These models are known as generative pre-trained transformers, which means they have been pre-trained to predict the next word in large amounts of text. Claude 3.5 Sonnet is the predecessor to the Claude 3 Sonnet introduced in March of this year. Claude 3.5 Sonnet is likely to be the middle model (based on parameter size) in the upcoming series of AI models by Anthropic — the smallest and biggest models are yet to be released. Anthropic has said Claude 3.5 Sonnet outperforms Claude 3 Opus by a huge margin. The new model is claimed to be twice as fast as the Claude 3 Sonnet. How does Claude 3.5 Sonnet perform? According to Anthropic, Claude 3.5 Sonnet sets some new industry benchmarks in capabilities such as coding proficiency (HumanEval), graduate-level reasoning (GPQA), and undergraduate-level knowledge (MMLU). The company claims that the new model has also shown significant improvement in grasping nuance, humour, and complex instructions. Claude 3.5 Sonnet is exceptional at writing high-quality content with a natural and relatable tone, according to Anthropic. Introducing Claude 3.5 Sonnet—our most intelligent model yet. This is the first release in our 3.5 model family. Sonnet now outperforms competitor models on key evaluations, at twice the speed of Claude 3 Opus and one-fifth the cost. Try it for free: pic.twitter.com/qz569rES18 — Anthropic (@AnthropicAI) June 20, 2024 Based on the benchmark scores shared by Anthropic on its official website, Claude 3.5 Sonnet seems outstanding. It has outdone GPT-4o, Gemini 1.5 Pro, and Meta’s Llama 3 400B in seven out of eight overall benchmarks. However, benchmark scores should not be taken too seriously — many AI startups have been accused of cherry-picking scores under categories that make them look good. What about Claude 3.5 Sonnet’s vision capabilities? Anthropic claims that Claude 3.5 Sonnet is its strongest vision model. A vision model in AI is a model capable of interpreting and analysing visual data such as images and videos. According to the company, the improvements in Claude 3.5 Sonnet are most noticeable for tasks that require visual reasoning such as decoding charts and graphs. The model is also capable of accurately transcribing text from imperfect images. For instance, The Indian Express clicked a random picture from Claude's iOS app and asked about the location. The model immediately identified the location by reading a poster and text on the distant wall. This ability to transcribe is what makes Claude 3.5 Sonnet beneficial for retail, logistics, and financial services, where AI may rely more on insights from an image, graphic, or illustration than from text, according to Anthropic.
