Chinese artificial intelligence start-up Moonshot AI don apologise after di roll-out of im new model, Kimi K3.1 Pro, stumble, and as promotional scrutiny around di launch dey widen, South China Morning Post report.
Di apology no come alone. Di company, wey be developer of di popular Kimi model line, also dey face one alleged paid influencer scandal wey dem dey talk about on X, according to VentureBeat. Di report say di matter don put di launch under serious scrutiny.
Kimi K3.1 Pro na 3-trillion parameter post-trained version of Kimi K3, di model wey Moonshot release back in July. Na on Friday di company unveil am, and e immediately begin set benchmark records. Moonshot and third-party AI model testing firm Artificial Analysis both score am high, and di results suggest say e go almost certainly take di crown as di top open weights AI model for di world when Moonshot release di weights under open license on November 4, 2026.
Before dat day reach, di model don already dethrone Xiaomi MiMo-V2.6-Pro, DeepSeek V4.1-Flash-1017, Meta new Muse Wonder, and even OpenAI GPT-6 Astra on Artificial Analysis Intelligence Index.
Wetin Kimi K3.1 Pro dey do well pass na complex coding, financial analysis, 3D creation and software-security work. Di report talk about large improvements for tasks wey need repeated coding, tool use, debugging and revision.
For now, developers and users fit access di model through im applications and developer application programming interface, and di rates na half of wetin Claude Opus 5.5 and Meta Muse Wonder dey charge for input and output. DeepSeek and Xiaomi still remain cheaper, though.
But di launch no waka without complaint. South China Morning Post note say na no be di first time di Chinese start-up dey face challenges serving im models because of computing constraints. Dat one join di promotional scrutiny wey surround di roll-out and di influencer allegations wey spread on X.
