NLP Lineage

15 July 2024 · Multimodality

Qwen2

Qwen2 Technical Report

A. Yang et al. (Qwen Team)

Date
15 July 2024
What the date marks
arXiv submission
Kind
Model
Organisation
Alibaba
Access
Apache 2.0 (72B custom)
Threads
In the guided tour
Chapter 8. The weights get out
Learning phases
6. Local models and deployment

A milestone: 1 later milestone descends directly from it, and its descendants reach 2 of the 10 threads. How levels are decided

Read the primary source (arxiv.org)

Full summary in progress. Until then, this page shows the milestone's record from the research data.

What it changed

Before it

Open weights meant Western labs, with English-centric training data.

The problem

Most of the world's users and much of its text are not English.

What it introduced

A full family from 0.5B to 72B with strong multilingual coverage, mostly Apache 2.0.

Why it matters

Qwen became the most downloaded and fine-tuned open family, and the base for a large share of derivative models.

What modern models inherit

A genuinely multilingual open baseline.

What came next

Qwen2.5, Qwen3 and the 2026 releases.

Before and after

The milestones it descends from, and the milestones that descend from it. Work developed alongside it is not counted. Weak links are left out unless you include them.

Before: where it came from

Qwen2 has 14 documented ancestors across 6 chapters, reaching back to Recurrent networks in 1990. Showing its direct parents and the 6 turning points among them.

Qwen2 has 14 documented ancestors across 6 chapters, reaching back to Recurrent networks in 1990. Showing all of them.

Qwen2 has 15 recorded ancestors, including indirect ones, across 6 chapters, reaching back to Recurrent networks in 1990. Showing its direct parents and the 6 turning points among them.

Qwen2 has 15 recorded ancestors, including indirect ones, across 6 chapters, reaching back to Recurrent networks in 1990. Showing all of them.

  1. 2024Qwen2This milestone.
  2. 2023Grouped-query attention Turning pointQwen2 built directly on thisChapter 9. Making it run anywhere
  3. 2021RoPE Turning pointQwen2 built directly on thisChapter 10. Reading more than you remember
  4. 2019Multi-query attentionGrouped-query attention built directly on thisChapter 9. Making it run anywhere
  5. 2019Transformer-XLRoPE was influenced by thisChapter 4. Pretrain once, use everywhere
  6. 2017Transformer Turning pointRoPE built directly on thisChapter 3. Attention, then only attention
  7. 2016Layer normTransformer built directly on thisChapter 3. Attention, then only attention
  8. 2015Residual connectionsTransformer was influenced by thisChapter 3. Attention, then only attention
  9. 2015Luong attentionTransformer was influenced by thisChapter 3. Attention, then only attention
  10. 2014AdamTransformer was influenced by thisChapter 2. Words become vectors
  11. 2014Seq2seq Turning pointTransformer built directly on thisChapter 2. Words become vectors
  12. 2014Attention Turning pointTransformer built directly on thisChapter 3. Attention, then only attention
  13. 2014GRU / encoder-decoderAttention built directly on thisChapter 2. Words become vectors
  14. 1997LSTM Turning pointSeq2seq built directly on thisChapter 1. Language as probability
  15. 1990Recurrent networksLSTM was influenced by thisChapter 1. Language as probability
  1. 2024Qwen2This milestone.
  2. 2023Grouped-query attention Turning pointQwen2 built directly on thisChapter 9. Making it run anywhere
  3. 2021RoPE Turning pointQwen2 built directly on thisChapter 10. Reading more than you remember
  4. 2019Multi-query attentionGrouped-query attention built directly on thisChapter 9. Making it run anywhere
  5. 2019Transformer-XLRoPE was influenced by thisChapter 4. Pretrain once, use everywhere
  6. 2017Transformer Turning pointRoPE built directly on thisChapter 3. Attention, then only attention
  7. 2017ConvS2STransformer was influenced by this (indirect influence)Chapter 3. Attention, then only attention
  8. 2016Layer normTransformer built directly on thisChapter 3. Attention, then only attention
  9. 2015Residual connectionsTransformer was influenced by thisChapter 3. Attention, then only attention
  10. 2015Luong attentionTransformer was influenced by thisChapter 3. Attention, then only attention
  11. 2014AdamTransformer was influenced by thisChapter 2. Words become vectors
  12. 2014Seq2seq Turning pointTransformer built directly on thisChapter 2. Words become vectors
  13. 2014Attention Turning pointTransformer built directly on thisChapter 3. Attention, then only attention
  14. 2014GRU / encoder-decoderAttention built directly on thisChapter 2. Words become vectors
  15. 1997LSTM Turning pointSeq2seq built directly on thisChapter 1. Language as probability
  16. 1990Recurrent networksLSTM was influenced by thisChapter 1. Language as probability

After: what it led to

Qwen2 has 3 documented descendants across 1 chapter, reaching forward to Qwen3.8 in 2026. Showing its direct descendants.

Qwen2 has 3 documented descendants across 1 chapter, reaching forward to Qwen3.8 in 2026. Showing all of them.

Qwen2 has 3 recorded descendants, including indirect ones, across 1 chapter, reaching forward to Qwen3.8 in 2026. Showing its direct descendants.

Qwen2 has 3 recorded descendants, including indirect ones, across 1 chapter, reaching forward to Qwen3.8 in 2026. Showing all of them.

  1. 2024Qwen2This milestone.
  2. 2025Qwen3built directly on Qwen2Chapter 8. The weights get out
  3. 2026Qwen3.5built directly on Qwen3Chapter 8. The weights get out
  4. 2026Qwen3.8built directly on Qwen3.5Chapter 8. The weights get out
  1. 2024Qwen2This milestone.
  2. 2025Qwen3built directly on Qwen2Chapter 8. The weights get out
  3. 2026Qwen3.5built directly on Qwen3Chapter 8. The weights get out
  4. 2026Qwen3.8built directly on Qwen3.5Chapter 8. The weights get out

Direct links and their sources

Where it came from

Built directly on

What it led to

Built on directly by

Terms