Best models under 4GB

Models whose smallest published quantization fits in under 4GB. These run on laptops, older cards, and integrated graphics.

From the file· live filter over real dataFrom the file· 40 models

How this is ranked

Ranked by downloads among models with a real quantization under 4GB. Size is measured from published file bytes, not estimated.

Best models under 4GB

#ModelParamsSmallest quantSmallest quant
1Qwen3.5-9B9.7B2.97 GiB2.97 GiB
2gemma-4-12B-it12.0B3.92 GiB3.92 GiB
3nemotron-3.5-asr-streaming-0.6b638M0.38 GiB0.38 GiB
4Qwen3.5-4B4.7B1.42 GiB1.42 GiB
5parakeet-unified-en-0.6b618M0.44 GiB0.44 GiB
6gemma-4-E4B-it8.0B3.30 GiB3.30 GiB
7Qwen3-4B4.0B1.01 GiB1.01 GiB
8FLUX.2-klein-9B9.1B3.71 GiB3.71 GiB
9Qwen3-8B8.2B2.12 GiB2.12 GiB
10cohere-transcribe-03-20262.1B1.41 GiB1.41 GiB
11HyperCLOVAX-SEED-Text-Instruct-1.5B1.6B1.06 GiB1.06 GiB
12Qwen3.5-0.8B873M0.31 GiB0.31 GiB
13Llama-3.2-1B-Instruct1.2B0.39 GiB0.39 GiB
14gemma-4-E2B-it5.1B2.13 GiB2.13 GiB
15parakeet-tdt-0.6b-v3627M0.39 GiB0.39 GiB
16Qwen3-VL-8B-Instruct-abliterated-v18.8B1.97 GiB1.97 GiB
17Llama-3.1-8B-Instruct8.0B2.02 GiB2.02 GiB
18ced-base86M0.12 GiB0.12 GiB
19Ace-Step1.5160M0.04 GiB0.04 GiB
20Qwen2.5-7B-Instruct7.6B2.59 GiB2.59 GiB
21Qwen3-TTS-12Hz-0.6B-Base915M0.50 GiB0.50 GiB
22Qwythos-9B-v29.7B3.64 GiB3.64 GiB
23UI-TARS-1.5-7B8.3B2.81 GiB2.81 GiB
24gemma-3-1b-it1000M0.52 GiB0.52 GiB
25gemma-4-E2B-it-qat-q4_0-unquantized5.1B3.12 GiB3.12 GiB
26Qwen3-1.7B2.0B0.50 GiB0.50 GiB
27whisper-medium764M0.25 GiB0.25 GiB
28embeddinggemma-300m303M0.26 GiB0.26 GiB
29Llama-3.2-3B-Instruct3.2B0.85 GiB0.85 GiB
30Qwen3-0.6B752M0.20 GiB0.20 GiB
31Qwen3-14B14.8B3.56 GiB3.56 GiB
32Voxtral-Mini-4B-Realtime-26024.4B2.35 GiB2.35 GiB
33Wan2.1-T2V-1.3B1.4B0.61 GiB0.61 GiB
34LFM2.5-1.2B-Instruct1.2B0.45 GiB0.45 GiB
35Qwen2.5-Coder-7B-Instruct7.6B2.59 GiB2.59 GiB
36Qwen3-VL-4B-Instruct4.4B1.01 GiB1.01 GiB
37Qwen2.5-1.5B-Instruct1.5B0.56 GiB0.56 GiB
38Jan-v3-4B-base-instruct4.4B1.56 GiB1.56 GiB
39gemma-3-4b-it4.3B1.43 GiB1.43 GiB
40whisper-large-v31.5B0.49 GiB0.49 GiB
Spec sheetFrom the fileFrom the filewhat these mean