Primary source@koltregaskes
View sourceEvent ·Multi-source
GPT-6 2.2 Sol 与 Astra 发布,Astra 在 ARC-3 上满分
OpenAI 发布 GPT-6 2.2 系列的两个模型:Sol 和 Astra。Astra 在 AGI ARC-3 基准上取得 100% 满分成绩,不过该基准此前已接近饱和,含金量存疑。Sol 则被认为相比前代没有明显升级。 值得注意的是,这两个模型在对比评测中都没有跑赢对应的 Claude 模型。这一结果延续了近期 Claude 系列在多类基准上的领先态势,也说明单看某个饱和基准的分数已难以区分模型实力。
Current read
OpenAI 这波更新有点平淡:Sol 没什么进步,Astra 虽然把 ARC-3 基准刷满了,但那基准本来就快饱和了。而且俩模型都没赢过 Claude。想选模型的人可以看看实测对比再决定。
11 sources61° AI heatLast updated 2026/11/17 18:30:03
Evidence
11 sourcesRelated source 1OpenAI
View sourceRelated source 2Artificial Analysis
View sourceRelated source 3ARC Prize
View sourceRelated source 4kimmonismus
View sourceRelated source 5The Rundown AI
View sourceRelated source 6TestingCatalog
View sourceRelated source 7宝玉
View sourceRelated source 8Greg Brockman
View sourceRelated source 9IT之家
View sourceRelated source 10Hunyuan
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.