airllm
AirLLM runs 70B large language models on a single 4GB GPU without quantization, distillation or pruning. Kimi K3 2.8T on under 4GB, Qwen3.8-Flash-Next 125B on 6GB, DeepSeek-V3 671B on ~12GB. Train Qwen3.8-Flash-Next under 6GB.
airllm has been downloaded 294,892 times in total on PyPI, including 7,493 in the last 30 days. The latest version is 4.0.0, released Sep 5, 2026.
Version4.0.0
Downloads
294.89k
License—
AuthorGavin Li
UpdatedSep 5, 2026
Downloads
Weekly, last 90d.
Includes CI traffic.
VersionsTotal4.*3.*
Range
View
Granularity
Group by
CI traffic
Stack: OffCI: Included3 / 40 series
Selected total79.51k
294.9kAll-time
7.5kLast 30 days
428Last 24 h
<0.01/sPer second
Sponsored
Sponsorships keep pepy free to read
Version distribution
Share of downloads by released version. Computed over the last quarter.
- 0132.0%
3.1.0
14.4k downloadsDownloads14.4k32.0% - 0217.9%
3.0.1
8.1k downloadsDownloads8.1k17.9% - 0310.4%
4.0.0
4.7k downloadsDownloads4.7k10.4% - 049.2%
3.2.0
4.1k downloadsDownloads4.1k9.2% - 054.9%
3.3.0
2.2k downloadsDownloads2.2k4.9% - 063.3%
2.11.0
1.5k downloadsDownloads1.5k3.3% - 071.5%
3.0.0
695 downloadsDownloads6951.5% - 080.8%
2.8.3
346 downloadsDownloads3460.8% - 0919.9%
Other
9.0k downloadsDownloads9.0k19.9%
Guess the next day
Thirteen recent days of airllm downloads. Drag the green handle on the right to guess where day fourteen lands.