Claudius Instant 1.2 is unveiled! It is designed with the aim of generating text very quickly, solving mathematical problems or tackling complex areas such as security or programming.
Claude artificial intelligence was created by Anthropic, a start-up co-founded by former OpenAI executives. It was designed as a competitor to other AI systems such as ChatGPT or Google Bard, and its newest update will significantly increase its performance. Claude Instant 1.2 is progressing quickly, very quickly!
A clear leap forward in performance
Anthropic hasn’t been idle in recent months, and Claude is showing test results that are clearly superior to the previous version. Even if the latter was already rather impressive, especially in terms of its responsiveness, the new Claude exceeded all his previous records. The gains are significant, especially with regard to programming and solving mathematical problems.
In a Python language programming benchmark, Claude 1.2 managed to achieve a score of 58.7%, while the older version scored 52.8%. This represents approximately 11.8% progress. For solving mathematical problems, Claude 1.1 reached 80.9%. Claude 1.2 surpasses it with a nice score of 86.7%, an improvement of almost 7.12%. Even though these margins may seem small, they are important to note, because they are a witness to increased security as well as a reduction in possible errors.
More accurate and more consistent
Claude 1.2 does not just improve its raw scores in the various tests. Anthropic also highlights its increased accuracy and improved consistency in the answers it gives. AI models sometimes generate false or inconsistent information. They’re commonly called “hallucinations,” and companies working on AI models are chasing them.
The company has also worked to strengthen Claude’s resistance to jailbreak attempts, which is a clear advantage in terms of security. Thus, the model’s security measures are much more difficult to circumvent, limiting the dangerous uses that could be made of them.
Gradually, Claude 1.2 is gaining ground in the field of generative AI models. Clearly, it is now a serious competitor in the artificial intelligence market. Anthropic is in any case on the right track if she continues to pamper her baby in this way.
0