aifollow.news
Back Rohan Paul
Rohan Paul· @rohanpaul_ai · X· · Original publication time AI score47

Distributed Cognition says VINCO beat several larger models at CatanMachine translation

AI introduction

Distributed Cognition says its 0.5 million-parameter VINCO beat Claude Opus 5.5, Claude Fable 5.1, Astra and Kimi K3 in 36 of 40 Catan games. Catan requires decisions without seeing an opponent’s cards or knowing the next dice roll; the match results are claims in the supplied account.

Original post · AI translation

@rohanpaul_ai
Translation · refer to the original

Distributed Cognition 的 VINCO 是一个拥有 50 万参数的模型,在卡坦岛桌游的 40 场对局中赢了 36 场,击败了 Claude Opus 5.5、Claude Fable 5.1、Astra 和 Kimi K3。

卡坦岛考验的是编程基准测试所不考验的一种能力:在看不到对手手牌、也不知道下一次掷骰结果的情况下选出好棋。

该公司更广泛的观点是,超级智能将来自许多像这样的模型相互协作,而不是由一个模型包揽一切。按该公司的数据,VINCO 的规模约为它在测试中击败的前沿大型语言模型的 1000 万分之一。

Quotedistributed cognition@distcog
Translation
我们很高兴推出 VINCO,这是我们训练来玩并赢得 1 对 1 卡坦岛对局的前沿 AI 模型。 我们已秘密将它部署到 https://t.co/jVgSFYCfkd 全球排行榜上;它曾登上第 1 名和第 2 名,并持续击败每一位人类世界冠军。
Original

We're excited to unveil VINCO, our frontier AI model trained to play and win 1v1 Catan. We've secretly deployed it on the https://t.co/jVgSFYCfkd global leaderboard, where it has reached both the #1 and #2 positions, consistently beating every human world champion.

View this post on X ↗

Source: Rohan Paul · x.com

Model
Found an error? Send a correction