Event ·Multi-source

Matt Shumer 用 4000 亿 Token 实测 Fable 5.2

Matt Shumer 宣布用超过 4000 亿 Token 测试 Fable 5.2,测试方式不是跑基准分数,而是让模型实际完成项目。他列出了其中 5 个测试案例,但帖子未披露具体案例内容和模型性能数据。该测试由 X 账号 FinanceYF5 转发,目前互动数据较低,原始评测细节有待后续公开。

Current read

Matt Shumer 拿 4000 亿 Token 让 Fable 5.2 真做项目而不是刷分,这种实测方式比看榜单更实在。原文还列出 5 个具体案例,做长上下文评测或者想换模型的人可以看看他怎么设计测试。

11 sources27° AI heatLast updated 2026/11/10 05:59:10

Evidence

11 sources
Related source 1Matt Shumer
View source
Related source 2Bindu Reddy
View source
Related source 5Mike Krieger
View source
Related source 6Alex Albert
View source
Related source 7lmarena.ai
View source
Related source 9kimmonismus
View source
Related source 10Liputan6 Tekno
View source

Conflict 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.