Primary sourcerohanpaul_ai
View sourceEvent ·Multi-source
Distributed Cognition 推出 VINCO:0.5M 参数模型在 Catan 中战胜 Claude
Distributed Cognition 发布 VINCO,一个仅 0.5M 参数的小模型。在 40 局卡坦岛(Catan)对局中,它赢了 36 局,击败对手包括 Claude Opus 5.5、Claude Fable 5.1、Astra 和 Kimi K3。按该公司统计,VINCO 的参数量比这些前沿 LLM 小约 1000 万倍。卡坦岛考验的是在不完全信息下做决策,这是编码基准测不出的能力。公司认为超级智能会来自多个这类专用模型协作,而非单一全能模型。
Current read
一个 0.5M 参数的小模型在卡坦岛赢了 40 局里的 36 局,对手是 Claude Opus 5.5 和 Kimi K3,体积还小了 1000 万倍,思路很反直觉。
11 sources63° AI heatLast updated 2026/11/5 21:27:19
Evidence
11 sourcesRelated source 1Artificial Analysis
View sourceRelated source 2Thomas Wolf
View sourceRelated source 3lmarena.ai
View sourceRelated source 4IT之家
View sourceRelated source 5shao__meng
View sourceRelated source 6SuperTechFans
View sourceRelated source 7@koltregaskes
View sourceRelated source 8Decoder
View sourceRelated source 9Allen AI (Ai2)
View sourceRelated source 10kimmonismus
View sourceConflict check
Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.