The race for the best artificial intelligence continues: Anthropic launches its latest model, Claude 3.5 Sonnet, which is reportedly on par with or superior to OpenAI's GPT-4 and Google's Gemini in many tasks. The new model is already available to Claude users on the web and iOS and will also be made available to developers.

Claude 3.5 Sonnet is ultimately intended to be the mid-tier model in the series – Anthropic uses the name Haiku for the smallest model, Sonnet for the mid-range, and Opus for the high-end model. According to the company, 3.5 Sonnet significantly outperforms the previous 3 Opus, with benchmarks showing a substantial margin of improvement. Furthermore, the new model is said to be twice as fast as the previous one, representing a significant advantage.

Benchmarks for AI models should always be viewed with caution as they quickly become outdated and can easily be selected to show the desired result. Nonetheless, Claude 3.5 Sonnet seems to be impressive: It outperformed GPT-4, Gemini 1.5 Pro, and Meta's Llama 3 400B in seven out of nine general benchmarks and four out of five vision benchmarks. This suggests that Anthropic has developed a legitimate competitor in this field.

What does that mean exactly? Anthropic says that Claude 3.5 is significantly better at writing and translating code, handling multi-step workflows, interpreting diagrams and graphics, as well as transcribing texts from images. The new Claude also understands humor better and can write in a much more human-like manner.

In addition to the new model, Anthropic introduces a new feature called Artifacts. With Artifacts, users can view and interactively edit the results of their Claude queries: for example, if the model designs something for the user, it can now show what it looks like, and the user can edit it directly in the app. This small but clever feature allows these AI tools to be more than just simple chatbots.

Artifacts also signal the long-term vision for Claude.

Here is the translation of the heading to English:

"At present, however, the new model is the big news. The speed of improvements is impressive: Anthropic launched Claude 3 Opus in March, boasting that it was as good as GPT-4 and Gemini 1.0, before OpenAI and Google released better versions of their models. Now Anthropic has made its next move, and it surely won’t be long before the competition follows suit. While Claude may not be mentioned as frequently as Gemini or ChatGPT, it is definitely in the running.