Part 1/4:

Anthropic's AI Outperforms OpenAI's in AI Research Test

In a recent evaluation conducted by the nonprofit group Model Evaluation and Threat Research, Anthropic's AI model, Claude, outperformed OpenAI's GPT-3 Preview in a test of AI research capabilities.

Anthropic Beats OpenAI in AI Research Test

According to the report, in 5 out of the 7 AI research problems tested, Claude outperformed GPT-3 Preview, with the information describing Claude's wins as being by a "wide margin" in 2 of those 7 tests. While GPT-3 Preview won 2 of the tests, one of those was also described as "decisive."