한국어EN
Open-Weight LineageAlibaba QwenAlibaba

Alibaba releases Qwen2.5

Made by AIWritten by AI (Claude) from the source below. Check anything important against the source.How it's made

Released Qwen2.5 — seven sizes from 0.5B to 72B, pretrained on up to 18T tokens, 128K context, 29+ languages; Qwen2.5-Coder and Qwen2.5-Math released alongside

Why it matters

All open models except the 3B and 72B were released under Apache 2.0, and the 14B and 32B sizes missing from the previous generation returned

Source — Primary sources first; secondary reports are labelled “Press”

Officialqwenlm.github.iohttps://qwenlm.github.io/blog/qwen2.5/

Added in Week 41, 2026 · See that week

On the map — ⑨ Open-Weight Lineage line

  1. Previous stopEXAONE 3.0
  2. This recordQwen2.5
  3. Next stopLlama 3.2
See it on the map

In the model family tree — Qwen family · 13 versions

  1. Earlier versionQwen2
  2. This recordQwen2.5
  3. Later versionQwQ-32B-Preview
See the whole Qwen family

Same week

Contextual RetrievalOfficialZayne Sprague et al. publish "To CoT or not to CoT?"PaperVoyage AI releases voyage-3OfficialKyutai unveils MoshiPaper

Nearby on the same axis

Meta releases Llama 3.2 with vision and on-device modelsOfficialUpstage releases Solar Pro PreviewOfficialGENIAC cycle 1 yields PLaMo-100B-Pretrained (100B parameters, from scratch)OfficialHermes 3Official