LLM WORLD

Qwen2.5-72B

Alibaba Qwen · · 104 days after Qwen2-72B

Verified

3 of 3 published values traced to a source2 primary sources, 2 archived

Why this status

Renamed from the series to the shipped flagship, which Alibaba describes as "a 72B-parameter dense decoder-only language model". The 72B and 3B variants are excluded from the series-wide Apache 2.0 licence.

What these states mean · How records are judged · Has this record changed?

The largest open Qwen2.5 model, released alongside the Qwen2.5-Coder and Qwen2.5-Math lines.

multilingualopen-weights

Details

ReleasedSeptember 19, 2024
stated in qwenlm.github.io, huggingface.co

Qwen2.5-72B was released on September 19, 2024.

CompanyAlibaba Qwen
FamilyQwen
Lineageafter Qwen2-72B · before Qwen3.6-35B-A3B
EraMultimodal & open weights
TypeLanguage model · LLM
ModalitiesNot recorded
CapabilitiesMultilingual
WeightsOpen weights
LicenceQwen License Agreement
Context window131K tokens
stated in huggingface.co

Qwen2.5-72B supports a context window of 131,072 tokens.

Parameters72B
stated in qwenlm.github.io, huggingface.co

Qwen2.5-72B has 72,000,000,000 parameters.

Current information, elsewhere

This record is a historical one: every figure above is traced to a dated primary source. For what is true today — live pricing, current rankings, availability — these are better maintained than anything here could be.

Sources

Record status: verified · confidence 90/100 · 2 of 2 primary

Renamed from the series to the shipped flagship, which Alibaba describes as "a 72B-parameter dense decoder-only language model". The 72B and 3B variants are excluded from the series-wide Apache 2.0 licence.

Primary means published by the organisation that made the model. Verified records require at least one.

Qwen family

  1. Qwen1.5-72B February 4, 2024
  2. Qwen2-72B June 7, 2024
  3. Qwen2.5-72B September 19, 2024
  4. QwQ-32B March 6, 2025
  5. Qwen3-235B-A22B April 29, 2025
  6. Qwen3-30B-A3B April 29, 2025
  7. Qwen3-Coder-480B-A35B-Instruct July 22, 2025
  8. Qwen3-Max-Thinking January 23, 2026
  9. Qwen3-Coder-Next February 2, 2026
  10. Qwen3.5-397B-A17B February 14, 2026
  11. Qwen3.6-Plus April 2, 2026
  12. Qwen3.6-35B-A3B April 15, 2026
  13. Qwen3.6-27B April 21, 2026
  14. Qwen3.7-Max May 16, 2026
  15. Qwen3.7-Plus May 21, 2026
  16. Qwen3.8-Max August 3, 2026
  17. Qwen3.8-Flash-Next August 3, 2026
  18. Qwen3.8-27B August 5, 2026
  19. Qwen3.8-2.4T-A95B August 8, 2026
  20. Qwen3.8-Omni-Flash September 18, 2026

What changed from Qwen2-72B

Compared with the previous release in this series. A field appears only when both records state a value — where one does not, it is listed as uncomparable rather than dropped, because a gap in our research is not a finding about the model.

Not comparable: Capabilities and modalities (not researched on both releases)

Compare them side by side →

Browse models

Compare Qwen2.5-72B with another model →
See Qwen2.5-72B on the interactive timeline →