@KevinNaughtonJr: Picking which tool an agent calls next is worth getting right. Subquadratic reports a 61.6 on Hugging Face's Decision I…
Summary
Subquadratic 发布 SubQ Snap，号称首个面向企业级真实工作负载的长上下文决策模型，在 Hugging Face Decision Index 上取得 61.6/61.7 的成绩，超过 Jev 的 57.9，且声称每 token 计算量不随上下文增长而增加。
View Cached Full Text
Cached at: 10/12/26, 07:30 PM
Picking which tool an agent calls next is worth getting right.
Subquadratic reports a 61.6 on Hugging Face’s Decision Index for SubQ Snap, ahead of Jev’s 57.9.
That’s a reason to test it on your own workflows.
Alexander Whedon (@alex_whedon): Introducing SubQ Snap, the first long-context decision model for real enterprise work.
Snap is the only decision model or LLM where compute per token stays constant as context grows.
• 384K tokens of multimodal context, 12x larger than Jev. • 61.7 on Decision Index v0.2.1 vs
Similar Articles
@rohanpaul_ai: Decision models deserve more attention than they get. look at the latency numbers on SubQ Snap from @subquadratic 660ms…
推文讨论 SubQ Snap 的延迟数据：100K tokens 约 660ms、384K tokens 约 2.5 秒，接近线性扩展而非二次注意力的 15 倍增长，使决策模型能一次处理超长多模态上下文。
@LinusEkenstam: 100,000 tokens processed in under ONE SECOND. What the actual hell. Snap just launched, and the numbers are ridiculous.…
SubQ 发布长上下文决策模型 Snap，支持 384K 多模态上下文、每 token 计算成本恒定，并以极低定价（$0.05/百万输入 token，输出免费）冲击企业级 AI 经济性。
Do you guys think subquadratic actually has a 12 million context model
Sub Quadratic claims to have a model with a context of 12 million tokens, but access is limited to partners; it performs well in the "needle in a haystack" test, but lacks evidence of general reasoning ability, raising doubts.
Aplomb 1: open-weights 5.3B decision model, 1M context, text/image/video/audio in one request, #1 among 4B models on the Decision Index
EmpirioLabs released Aplomb 1, an open-weights 5.3B decision model built on Qwen3.5-4B that handles 1M-token text, image, video and audio inputs, returns probabilities for tool arguments, and tops the Decision Index board among sub-4B models while claiming leading scores on 8 of 38 benchmarks up to 5.3B.
@akshay_pachaar: https://x.com/akshay_pachaar/status/2109678653974155459
文章介绍了 TypeSafe AI 的 Jev 'System One' 决策模型在开源 Agent Harness 中的表现：在 40 个决策任务上达到与 Opus 相同的准确率，同时决策成本降低约 150 倍，并通过置信度校准将不确定答案路由到人工审核。