Industry

两周内15家厂商密集发布"决策模型",Vercel报告称26%请求已由其路由

Why we picked it

有人整理了一份完整时间线,两周内 Together AI、OpenAI、Google 等15家全上了自己的决策模型,连命名都互相抄。Vercel 那个26%路由占比数据挺值得琢磨,做 agent 架构的可以对照各家参数量和发布节奏看看选型。

这篇推文盘点了9月23日至11月11日期间15家公司先后推出"decision model"类产品的密集节奏:Together AI 的 Tev1-4B(9月23日)打头,随后 Alibaba Cloud 的 decision-model(9月24日)、Liquid AI 的 d1(9月29日)、Databricks 的 ai_decide(9月30日)、AWS 的 Strands Decider 2B 和 Cloudflare 的 Clef/Clef-flash(10月1日)、OpenAI 的 Decisions API(GPT-6 Luna,10月6日)、Microsoft-Decision-1(10月9日)、IBM Granite Switch 4.1 Decision 适配版(10月15日)、Google DeepMind 的 Gemini Decision 1(11月11日)等接连上线,其中 Clef-omni 和 Perplexity Decider v1.3 还发布了后续更新版本。

Image · Sebastian Raschka
Body

这类模型的核心功能是做路由决策,即判断请求应该交给哪个更大的模型处理。推文引用 Vercel 的报告称,目前 26% 的请求已经通过 decision model 路由,说明这一品类在几周内从零散尝试变成了行业标配。从 4B 级小模型(Tev1-4B、Strands Decider 2B)到 API 形态(OpenAI Decisions API),各家实现路径不一,但定位高度雷同。

名单覆盖美国(OpenAI、AWS、Google DeepMind、Microsoft)、中国(Alibaba Cloud)和欧洲(Liquid AI、Cloudflare)厂商,显示路由层竞争不分地域。对开发者而言,这类模型通常以 2B-4B 的小参数量换取低延迟和低成本,适合嵌入 agent 和网关架构作为分发前置层。

Original · Sebastian Raschka

Wow that escalated quickly. Was gone for a few weeks and now pretty much everyone launched their own decision model copycats:

- Sep 23: Together AI, Tev1-4B - Sep 24: Alibaba Cloud, decision-model - Sep 29: Liquid AI, d1 - Sep 30: Databricks, ai_decide - Oct 1: Amazon AWS, Strands Decider 2B - Oct 1: Cloudflare, Clef and Clef-flash (Clef-omni on Oct 9) - Oct 1: Perplexity, Decider (v1.3 on Nov 12) - Oct 6: OpenAI, Decisions API (GPT-6 Luna) - Oct 7: https://t.co/s8O5rwbq5z, H2O-Lightning-4B - Oct 7: Snowflake, Snowflake Decision - Oct 9: Microsoft, Microsoft-Decision-1 - Oct 15: IBM, Granite Switch 4.1 with Decision adapter - Oct 21: Baseten, Mammoth-1 and Elephant-1 - Nov 6: Salesforce, Koa Decisions - Nov 11: Google DeepMind, Gemini Decision 1

According to a Vercel report, 26% of requests are now routed through decision models 📈

Industry Coordinates & Source Verification

This report is categorized under AITOP's "Industry" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows and US-China foundation model benchmarks and commercialization comparison. Key entities and technical topics under continuous tracking include decision model, model routing, OpenAI, Gemini.

Source attribution: curated from Sebastian Raschka (@rasbt).