Awani Review

Complete News World

ChatGPT defeated?  OpenAI’s big competition is releasing faster and cheaper AI

ChatGPT defeated? OpenAI’s big competition is releasing faster and cheaper AI

Anthropic launches the new version It’s a competitor to ChatGPT. This is intended to be a cheaper alternative to ChatGPT, but no less efficient.

Is this the beginning of the confrontation between mankind And Open AI ? Remember that the first was founded by former members of the second. So they are keen to present a serious competitor to the very popular ChatGPT of their previous company. Announce Anthropic AI Claude Instant 1.2.2 Update It is faster but at an affordable cost.

Improved update across the board

Version 1.2 is therefore an update to the AI Claude Instant 1.1.2 Update. According to Anthropic, the new version of the ChatGPT competitor incorporates the strengths of Claudius 2. This is another leading artificial intelligence of OpenAI competitor.

The Claude Instant update shows significant improvements in areas such as math, cryptography, reasoning, and security. Internal tests clearly show the superiority of the new version. In a similar encryption test, Cloud Instant achieved a 1.2a score of 58.7% against only 52.8% for the previous version.

A series of mathematical problems was used to choose between the two models. The old model has arrived score of 80.9%while the update was better with a Score 86.7%.

ChatGPT’s least error-prone competitor

New version answers are More developed and better organized, according to Anthropic. Better design To adhere to the format guidelines will be the explanation. On the other hand, Model 1.2 improves citation extraction. Its multilingual capabilities have received significant improvements.

Anthropic also announced an update that brings more reliability. This is the one Less prone to hallucinations And More resistant to jailbreak attempts according to its designer.

Remember, in the language of artificial intelligence, “hallucinations” means generating incorrect or absurd text. For “jailbreaking”, the term refers to manipulations to circumvent the parameters of the main language model.

Less caliber than GPT-4

Anthropic aims to create a new generation of algorithms forSelf-learning artificial intelligence. Note that these algorithms can be used to create multitasking virtual assistants. The latter is effectively capable of replying to emails, performing searches and creating illustrations. GPT-4 He recently gave us a glimpse into these capabilities.

However, Cloud Instant is not such an algorithm. Instead, the model only competes with entry-level OpenAI solutions.

Either way, Anthropic says it now has thousands of customers and partners for its solutions. Cite a startup in california Quora among its clients. The Big Computer Forum makes it possible to take advantage of Claude and Claude Instant via its site chatbot bo. Another prestigious client, DuckDuckGo It is a search engine that aims to preserve the privacy of its users. His new virtual assistant Duck Assist Merge Claude.

See also  The French shadow displays new ambitions