OpenAI has announced the release of its flagship language model, named GPT-6 Astra. This new version demonstrates superiority over the previous model.
GPT-6 Astra holds leading positions in most tests concerning computer operation and cybersecurity. Furthermore, a significant reduction in the frequency of factual errors in generated responses is noted.
The model has become more reliable as it is less prone to hallucinations and practically does not violate established limitations. More detailed information about this model is available in its system card on the company's official website.