Moonshot K3.1 Pro High-Speed Version 2x to 3x Faster

On October 24, according to Jin10 data, Moonshot AI launched Kimi K3.1 Pro and its high-speed version, K3.1 Pro Highspeed, with the latter running 2 to 3 times faster than the standard version.

The official team stated that both models support a million-token context length, making them suitable for complete software project development, game production, 3D modeling, and financial analysis; in coding tests, the new models scored 65.7%, exceeding GPT-6 Astra's 58.2%, and achieved a 92.7% score in 3D modeling tests, nearing top closed-source models.

The official team also noted that in financial analysis, the new models can process financial data, cross-reference documents, and conduct investment analysis, with performance in enterprise operations tests approximately 4.2 times that of the previous generation; the hallucination rate per statement in specialized tests dropped from 23.4% to 8.3%. Both models are now available on Kimi.ai, Kimi Work, Kimi Code, and via API, with full model weights scheduled to be released before November 4.