首批家族包含 Qwen4-Max、Qwen4-Plus、Qwen4-Flash 以及开源主力 Qwen4-27B 四款产品。
This is a AI post classified by Jev as Frontier models (a model release), kept by the AI Radar because it carries real work, not commentary.
阿里的Qwen4 已经在训练中,即将正式发布: 首批家族包含 Qwen4-Max、Qwen4-Plus、Qwen4-Flash 以及开源主力 Qwen4-27B 四款产品。 它的核心定位是面向原生 Agent 时代设计的新架构(原生支持多工具复杂调度、长时间长上下文任务拆解与低激活比 MoE)。 旗舰版 Qwen4-Max 的总参数规模 总参数:约在 3T(3万亿)左右的超大规模 MoE 架构。 激活参数(Active Parameters):控制在 100B 左右。 不过我更看好他们的后续模型。阿里已经明确表示 Qwen4.5、Qwen5 等后续模型会继续扩大到 5–10T 参数规模。 Scaling Law还没有结束,至少阿里的判断依然是:更大的模型、更大的训练集群、更长的训练周期,仍然可以换来更强的智能。
Posted by lifcc (9.6k followers) 1 h ago · 7 likes · 1.2k views · view the original post on X. Kept by the AI Radar as Frontier models.
More AI work like this
- Confirmed by Alibaba: Qwen 4 27B is coming — @AestherML
- 🚨 OPUS 5.5 Leaks — @Priyannkaaaa
- Grok4.7第一波实测来了,Grok 4.7 (high in Grok build)和GPT-5.6 Sol(high in codex) — @xueyu1125
- Grok4.7第一波实测来了,Grok 4.7 (high in Grok build)和GPT-5.6 Sol(high in codex) — @xueyu1125
- Step 5 Preview 在真实 API 计价下处在 AA Intelligence (图1)和 Terminal-Bench 4(图2) 的帕累托前沿! — @Fei2411
- 🥊Grok 4.7 vs Xiaomi MiMo 2.6 Pro !🥊 — @YouWareAI
- Grok 4.7: giant. Gemini 4 Pro: giant. GPT-6 Astra: giant. Claude Fable 5.1: a pixelated… — @ZentrixHQ
- I built this browser-based drum machine / 16-step sequencer in about 15 minutes using… — @ChrisGPT
Every post is read and classified by Jev (TypeSafe): what it is, which market it belongs to, and whether the link is a real tool. 41.1k posts from 5k X accounts over the last 14 days, 1.7k tools, 19 markets. Collected every 5 minutes, fully re-ranked every hour — last update 2026-09-22 08:01 UTC. Full method.