Models

Thinky Machines 发布 Inkling 2 Small,ARC-AGI 基准成绩公开

Why we picked it

Thinky Machines 的 Inkling 2 Small 跑分出炉了:ARC-AGI-1 拿 96.5%、ARC-AGI-2 拿 81.7%,都压过 DeepSeek V4.1 Flash 1017。但 ARC-AGI-3 只有 2.6%,被 DeepSeek 的 5.8% 反超。关注小模型推理能力上限的开发者可以看下这份对比数据。

Thinky Machines 的小型模型 Inkling 2 Small 在 ARC Prize 官方评测中公布成绩:ARC-AGI-3 上得分 2.6%,通过 provider adapter 模式总花费 704 美元,standard 模式得分 0.3%、花费 858 美元。

Body

与前代模型的对比数据显示,Inkling 2 Small 在 ARC-AGI-1 上拿到 96.5%,超过 DeepSeek V4.1 Flash 1017 的 94.5%;在 ARC-AGI-2 上以 81.7% 领先对手的 78.1%。但在 ARC-AGI-3 上以 2.6% 落后于 DeepSeek V4.1 Flash 1017 的 5.8%。

ARC-AGI 系列基准专测抽象推理与泛化能力,ARC-AGI-3 难度最高,目前所有模型得分都很低。Inkling 2 Small 在低难度分项的接近满分表现说明小型模型也能在受限推理任务上逼近上限,但面对 ARC-AGI-3 这类需要长程探索的任务仍有明显差距。

Original · ARC Prize

Inkling 2 Small from @thinkymachines on ARC-AGI (Verified):

- ARC-AGI-3: 2.6%, $704 total (provider adapter); 0.3%, $858 total (standard)

It beat DeepSeek V4.1 Flash 1017 on ARC-AGI-1 (96.5% vs 94.5%) and ARC-AGI-2 (81.7% vs 78.1%), but trailed on ARC-AGI-3 (2.6% vs 5.8%). https://t.co/UJNieTlm9a

Industry Coordinates & Source Verification

This report is categorized under AITOP's "Models" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to US-China foundation model benchmarks and commercialization comparison and AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include Inkling 2 Small, Thinky Machines, DeepSeek, ARC-AGI.

Source attribution: curated from ARC Prize (@arcprize).