Anthropic launches groundbreaking AI model: faster, smarter, more cost-effective

2023-08-11

Since the release of ChatGPT by OpenAI, many companies have been trying to create their own AI models, but only a few have been able to stand out. Anthropic is one of them.

This artificial intelligence startup released their own AI model called Claude in March. It has proven to be a strong competitor to OpenAI's GPT-3.5 and GPT-4. Meanwhile, Anthropic also released Claude Instant, which, according to Anthropic, is a lighter, cheaper, and faster version of Claude. Now, it is being upgraded.

On Wednesday, Anthropic released Claude Instant 1.2, an improved version of the model that leverages the latest version of Claude, Claude 2.0, released in July.

According to the announcement, Claude Instant 1.2 has significant improvements in mathematics, coding, reasoning, and security due to the advanced features of Claude 2.0. It also generates longer and more coherent answers.

To test this model, Anthropic compared Claude Instant 1.1 and 1.2 in standard benchmark evaluations, including Codex evaluation and Grade-school math problem benchmark (GSM8k), which are good benchmarks for mathematical and coding abilities.

In both cases, the 1.2 version scored 58.7% in the Codex evaluation, while the original version scored 52.8%. In the GSM8k test, the 1.2 version scored 86.7%, while the original version scored 80.9%.

In other benchmark exams, the performance of the new version of the model was either slightly lower or slightly higher than the old version, with minimal differences.

The quality of the answer outputs has also been improved, with reduced hallucinations and increased resistance to jailbreaking attempts. A red-teaming evaluation found that Claude 1.2 is the most secure model to use.

RECENT AI NEWS

RECENT AI TOOLS