Kimi K3.1 Pro from Moonshot AI

Chinese company Moonshot AI introduced its flagship language model Kimi K3.1 Pro, positioned as a direct competitor to Claude Opus and GPT-6 Astra.

The new architecture boasts an impressive capacity of 2.8 trillion parameters and supports a context window of up to one million tokens.

In independent Humanity’s Last Exam tests, the model demonstrated leading positions, surpassing existing analogs in key intelligence metrics.

Developers announced the release of an open version with accessible model weights on November 4, allowing researchers to test the system's capabilities independently.

You can test the novelty through the official Moonshot AI platform.