AImodelsComparison.com 1,226 open · 200 API
Menu
Models Best of Trending Can I run it? Methodology About
1,226 models · 1,226 open weights · 200 API-only · refreshed nightly

Open-weights models

Models whose weights are published on Hugging Face — download them and run them on your own hardware. Many are also served by hosted providers; those carry an "Also via API" badge with live prices.

Type
All 1,426 Open weights (self-host) 1,226 API-only 200
Task
All text-generation 534 text-to-image 155 image-text-to-text 112 ASR 86 sentence-similarity 64 translation 56 text-classification 51 feature-extraction 48 image-classification 39 token-classification 21 text-to-speech 19 audio-classification 12 image-segmentation 11 summarization 8 object-detection 8 audio-text-to-text 1 any-to-any 1
Runs on
Any RTX 3060 · 12 GB RTX 4070 · 16 GB RTX 4090 · 24 GB 2 × 24 GB Mac M3 · 24 GB Mac M4 Max · 64 GB
Licence
Size
$
56 models
#ModelParamsContextCommercial use30dMin VRAM
01 t5-small google-t5 · translation Open weights 60 M — ✓ apache-2.0 23.9 M from 0.8 GB
02 t5-base google-t5 · translation Open weights 220 M — ✓ apache-2.0 1.9 M from 1.5 GB
03 opus-mt-nl-en Helsinki-NLP · translation Open weights — 512 ✓ apache-2.0 1.2 M —
04 nllb-200-distilled-600M facebook · translation Open weights 600 M 1 K ✗ cc-by-nc-4.0 889 K from 1.9 GB
05 opus-mt-fr-en Helsinki-NLP · translation Open weights 80 M 512 ✓ apache-2.0 611 K from 0.8 GB
06 nllb-200-distilled-1.3B facebook · translation Open weights 1.3 B 1 K ✗ cc-by-nc-4.0 482 K from 3.6 GB
07 opus-mt-zh-en Helsinki-NLP · translation Open weights — 512 ✓ cc-by-4.0 390 K —
08 opus-mt-de-en Helsinki-NLP · translation Open weights — 512 ✓ apache-2.0 282 K —
09 opus-mt-en-ru Helsinki-NLP · translation Open weights — 512 ✓ apache-2.0 247 K —
10 nllb-200-3.3B facebook · translation Open weights 3.3 B 1 K ✗ cc-by-nc-4.0 243 K from 8.3 GB
11 indictrans2-en-indic-dist-200M ai4bharat · translation · gated Open weights 275 M — ✓ mit 231 K from 1.7 GB
12 opus-mt-en-de Helsinki-NLP · translation Open weights — 512 ✓ cc-by-4.0 222 K —
13 t5-3b google-t5 · translation Open weights 2.9 B — ✓ apache-2.0 221 K from 13.5 GB
14 t5-large google-t5 · translation Open weights 740 M — ✓ apache-2.0 198 K from 3.9 GB
15 opus-mt-en-fr Helsinki-NLP · translation Open weights — 512 ✓ apache-2.0 180 K —
16 opus-mt-es-en Helsinki-NLP · translation Open weights — 512 ✓ apache-2.0 151 K —
17 opus-mt-en-es Helsinki-NLP · translation Open weights — 512 ✓ apache-2.0 135 K —
18 opus-mt-ru-en Helsinki-NLP · translation Open weights — 512 ✓ cc-by-4.0 121 K —
19 opus-mt-it-en Helsinki-NLP · translation Open weights — 512 ✓ apache-2.0 109 K —
20 opus-mt-tc-big-ko-en Helsinki-NLP · translation Open weights 210 M 1 K ✓ cc-by-4.0 102 K from 1.0 GB
21 opus-mt-ar-en Helsinki-NLP · translation Open weights — 512 ✓ apache-2.0 98 K —
22 Hy-MT2-30B-A3B tencent · translation Open weights Also via API 30.1 B 262 K ✓ apache-2.0 95 K from 71.2 GB
23 opus-mt-ko-en Helsinki-NLP · translation Open weights — 512 ✓ apache-2.0 95 K —
24 envit5-translation VietAI · translation Open weights — — ⚠ openrail 89 K —
25 opus-mt-en-zh Helsinki-NLP · translation Open weights — 512 ✓ apache-2.0 76 K —
26 opus-mt-ROMANCE-en Helsinki-NLP · translation Open weights — 512 ✓ apache-2.0 76 K —
27 opus-mt-tc-big-en-tr Helsinki-NLP · translation Open weights — 1 K ✓ cc-by-4.0 74 K —
28 opus-mt-en-az Helsinki-NLP · translation Open weights — 512 ✓ apache-2.0 68 K —
29 indictrans2-en-indic-1B ai4bharat · translation · gated Open weights 1.1 B — ✓ mit 66 K from 5.6 GB
30 sarvam-translate sarvamai · translation Open weights 4.3 B — gpl-3.0 61 K from 10.6 GB
31 opus-mt-fr-es Helsinki-NLP · translation Open weights — 512 ✓ apache-2.0 57 K —
32 opus-mt-it-es Helsinki-NLP · translation Open weights — 512 ✓ apache-2.0 55 K —
33 opus-mt-eu-es Helsinki-NLP · translation Open weights — 512 ✓ apache-2.0 55 K —
34 Index-Translate-9B-GGUF IndexTeam · translation Open weights 9.0 B — ✓ apache-2.0 50 K from 6.2 GB
35 Sugoi-32B-Ultra-GGUF sugoitoolkit · translation Open weights — — ✓ apache-2.0 50 K from 14.0 GB
36 opus-mt-sv-en Helsinki-NLP · translation Open weights — 512 ✓ apache-2.0 50 K —
37 TowerInstruct-13B-v0.1 Unbabel · translation Open weights 13.0 B 4 K ✗ cc-by-nc-4.0 49 K from 59.7 GB
38 Hunyuan-MT-7B-GGUF Mungert · translation Open weights — — unknown 47 K from 3.1 GB
39 sanskrit-en-custom-transformer krpraveen · translation Open weights — — ✓ mit 47 K —
40 Sugoi-14B-Ultra-GGUF sugoitoolkit · translation Open weights — — ✓ apache-2.0 46 K from 6.8 GB
41 opus-mt-de-fr Helsinki-NLP · translation Open weights — 512 ✓ apache-2.0 46 K —
42 nllb-200-1.3B facebook · translation Open weights 1.3 B 1 K ✗ cc-by-nc-4.0 44 K from 3.6 GB
43 mbart-large-50-many-to-many-mmt facebook · translation Open weights 610 M 1 K unknown 43 K from 3.3 GB
44 opus-mt-fr-de Helsinki-NLP · translation Open weights — 512 ✓ apache-2.0 41 K —
45 opus-mt-mul-en Helsinki-NLP · translation Open weights — 512 ✓ apache-2.0 37 K —
46 madlad400-3b-mt google · translation Open weights 2.9 B — ✓ apache-2.0 36 K from 2.8 GB
47 opus-mt-tc-big-ar-en Helsinki-NLP · translation Open weights — 1 K ✓ cc-by-4.0 35 K —
48 Hy-MT2-30B-A3B-GGUF tencent · translation Open weights 30.0 B — ✓ apache-2.0 35 K from 25.1 GB
49 opus-mt-en-ar Helsinki-NLP · translation Open weights — 512 ✓ apache-2.0 30 K —
50 opus-mt-pl-en Helsinki-NLP · translation Open weights — 512 ✓ apache-2.0 29 K —
VRAM figures are estimates for the smallest available quantization at 8K context — see /methodology. Downloads refresh nightly from the Hugging Face API; API prices (blended 3:1 input:output per 1M tokens) and Artificial Analysis indices from OpenRouter.