Kimi-K3 tops Arena's AI coding leaderboard with 1,679 points, pushing Anthropic's Claude Fable 5 into second place.

Kimi K3 topped Arena's coding leaderboard over Claude Fable 5 with 1,679 points. But a 51% hallucination rate shows why ...

Moonshot AI's 2.8-trillion-parameter model Kimi K3 tops Anthropic's Claude Fable 5 and OpenAI's GPT 5.6 Sol on impressive ...

Independent evaluators are pegging the Chinese, open-source Kimi K3 higher than anything the top US labs can offer. Kimi K3 has debuted ...

Moonshot AI's Kimi K3 scored 1,679 in Frontend Code Arena, beating Claude Fable 5 and GPT-5.6 Sol. Here's what it means for ...

LOL. This isn't really how metrics work. Or at least you just have to admit that it's a pure popularity contest unmoored from even a shared general concept of usefulness. I guess that's the invocation ...

OpenAI's GPT-5.6 Sol surpasses Anthropic's Claude Fable 5 in the Design Arena benchmark, marking a significant achievement ...

Beijing-based Moonshot AI has released Kimi K3, a 2.8 trillion parameter model that the company describes in its technical ...

Interesting approach for them to leverage the Elo rating system used in chess (and other) ratings. Seems like a clever way to rank these engines, while preserving a lot of the data that goes to the ...