lyogavin

airllm

Layer-by-layer inference that runs 70B models on a 4GB GPU — no quantization required; 405B on 8GB, DeepSeek-V3 671B on ~12GB. One AutoModel line for most open model families.

22,742 2,611 Jupyter Notebook Apache-2.0updated 4 days ago

local

View on GitHub Dispute this mapping →

Curator's take

The trick is elegant and the tradeoff is brutal, and you should know both: only one layer lives on the GPU at a time, so VRAM scales with layer size instead of model size — that's how 671B fits on a hobbyist card — but every token streams the whole model from disk, so generation runs at seconds-per-token. Use it for batch/offline jobs where 'it fits' beats 'it's fast', for poking at frontier-scale open models on hardware you own, or with block-wise 4/8-bit compression for a ~3x claw-back. NOT a chat or serving solution: Ollama is the fits-in-VRAM daily driver, vLLM the throughput server. Apple-silicon Macs supported via MLX. README carries the author's sponsor/affiliate links — the library stands on its own.

Mapped by ShipWithAI editors · links verified

README.md

airllm_logo

Quickstart | Configurations | MacOS | Example notebooks | FAQ

AirLLM dramatically reduces inference memory usage, letting 70B large language models run on a single 4GB GPU card — without quantization, distillation, or pruning. You can even run 405B Llama 3.1 on 8GB, and DeepSeek-V3 (671B) on ~12GB.

AI Agents Recommendation:

Updates

[2026/06] v3.0: FP8 model support + the latest models. Run DeepSeek-V3 (671B) on ~12GB and Qwen3-235B on ~3GB, plus Qwen3, Llama 3.x/4, DeepSeek V2/V3, Phi-4, Gemma and more — all through a single AutoModel.

[2024/08/20] v2.11.0: Support Qwen2.5

[2024/08/18] v2.10.1 Support CPU inference. Support non sharded models. Thanks @NavodPeiris for the great work!

[2024/07/30] Support Llama3.1 405B (example notebook). Support 8bit/4bit quantization.

[2024/04/20] AirLLM supports Llama3 natively already. Run Llama3 70B on 4GB single GPU.

[2023/12/25] v2.8.2: Support MacOS running 70B large language models.

[2023/12/20] v2.7: Support AirLLMMixtral.

[2023/12/20] v2.6: Added AutoModel, automatically detect model type, no need to provide model class to initialize model.

[2023/12/18] v2.5: added prefetching to overlap the model loading and compute. 10% speed improvement.

[2023/12/03] added support of ChatGLM, QWen, Baichuan, Mistral, InternLM!

[2023/12/02] added support for safetensors. Now support all top 10 models in open llm leaderboard.

[2023/12/01] airllm 2.0. Support compressions: 3x run time speed up!

[2023/11/20] airllm Initial version!

Star History

Quick start
Model Compression
Configurations
[Run on MacOS](#

Continue your stack

What teams reach for next — and why each earns a place beside airllm. Ranked by curator confidence.

Alternative to See all + compare →

Ollama

Run Llama, Mistral and other open models locally with a single command and a clean API.

Why it fitsBoth run open models on your own hardware, on opposite sides of one constraint: Ollama gives fast, polished local inference for models that fit your VRAM; AirLLM runs models that don't fit at all — 70B on 4GB — by streaming one layer at a time, at heavy latency cost.

vLLM

High-throughput, memory-efficient inference and serving engine for LLMs.

Why it fitsOpposite ends of the local-inference spectrum: vLLM maximizes throughput given abundant VRAM (production serving); AirLLM minimizes VRAM given abundant patience (frontier-size models on consumer cards).

airllm

AI Agents Recommendation:

Updates

Star History

Table of Contents

Continue your stack