Skip to content
AI Atlas
Open model frontier

Open models

Every canonical model whose weights can be downloaded, described by measurable properties — licence permissions, parameters, context, release, benchmark ranks, providers, estimated local fit. Openness is a set of observed dimensions here, never a score.

950 downloadable models

Openness explorer

Categories and licence permissions

Left: the downloadable universe by openness category. Right: what the licences on this page allow, counted per category — allowed · restricted · unknown.

Top licences: Apache License 2.0 (275) · Other (unclassified licence) (115) · MIT License (92) · Creative Commons Attribution-NonCommercial 4.0 (22) · Apple Sample Code / ML Research License (15) · Llama 3.2 Community License (12)

Category · this pageCommercial useRedistributionDerivativesHosting
Open source n=0
Open weights n=4524021240212402123022
restricted-weights n=5410500500230

Counts read: allowed · restricted · unknown / unclassified — over the 50 models on this page, from each licence's stated terms (ontology).

Downloadable models

950 models · sorted by best benchmark rank

Permissions, in order:commercial use· redistribution· derivatives· hosting allowed restricted unknown

ModelLicence · permissionsParamsContextReleaseBest resultsProvidersFit 64 GB @4bit · 128 GB @8bit Estimated
sdxl-turboOpen weightsStability AIOther2.57B27 Nov 20230✓ fits2.0 GB4bit✓ fits3.5 GB8bit
sdxl-vaeOpen weightsStability AIMIT83.7M21 Jun 20230✓ fits0.6 GB4bit✓ fits0.6 GB8bit
siglip-base-patch16-224Open weightsGoogleApache-2.0203.2M30 Sept 20230✓ fits0.6 GB4bit✓ fits0.7 GB8bit
siglip-base-patch16-512Open weightsGoogleunclassified0
siglip-so400m-patch14-384Open weightsGoogleApache-2.0878M8 Jan 20240✓ fits1.0 GB4bit✓ fits1.5 GB8bit
siglip2-base-patch16-224Open weightsGoogleApache-2.0375.2M17 Feb 20250✓ fits0.7 GB4bit✓ fits0.9 GB8bit
siglip2-giant-opt-patch16-384Open weightsGoogleApache-2.01.87B17 Feb 20250✓ fits1.6 GB4bit✓ fits2.6 GB8bit
siglip2-so400m-patch16-512Open weightsGoogleunclassified0
specter2_baseOpen weightsAllen Institute for AIApache-2.016 Feb 20230
stable-audio-3-mediumOpen weightsStability AIOther2.31B17 May 20260✓ fits1.8 GB4bit✓ fits3.2 GB8bit
stable-audio-3-medium-baseOpen weightsStability AIunclassified0
stable-audio-3-small-musicOpen weightsStability AIOther567.6M17 May 20260✓ fits0.8 GB4bit✓ fits1.2 GB8bit
stable-audio-3-small-music-baseOpen weightsStability AIunclassified0
stable-audio-3-small-sfxOpen weightsStability AIOther567.6M17 May 20260✓ fits0.8 GB4bit✓ fits1.2 GB8bit
stable-audio-3-small-sfx-baseOpen weightsStability AIunclassified0
stable-audio-open-1.0Open weightsStability AIOther1.21B24 May 20240✓ fits1.2 GB4bit✓ fits1.9 GB8bit
stable-diffusion-3-medium-diffusersOpen weightsStability AIOther2.08B12 Jun 20240✓ fits1.7 GB4bit✓ fits2.9 GB8bit
stable-diffusion-3.5-largeOpen weightsStability AIOther8.15B22 Oct 20240✓ fits5.2 GB4bit✓ fits9.9 GB8bit
stable-diffusion-3.5-mediumOpen weightsStability AIOther2.47B29 Oct 20240✓ fits1.9 GB4bit✓ fits3.3 GB8bit
stable-diffusion-xl-base-1.0restricted-weightsStability AIOpenRAIL++-M2.57B25 Jul 20230✓ fits2.0 GB4bit✓ fits3.5 GB8bit
stable-diffusion-xl-refiner-1.0restricted-weightsStability AIOpenRAIL++-M2.26B26 Jul 20230✓ fits1.8 GB4bit✓ fits3.1 GB8bit
stable-fast-3dOpen weightsStability AIOther1.01B17 Jul 20240✓ fits1.1 GB4bit✓ fits1.7 GB8bit
stable-video-diffusion-img2vidOpen weightsStability AIOther1.52B20 Nov 20230✓ fits1.4 GB4bit✓ fits2.3 GB8bit
stable-video-diffusion-img2vid-xtOpen weightsStability AIOther1.52B20 Nov 20230✓ fits1.4 GB4bit✓ fits2.3 GB8bit
stablelm-3b-4e1tOpen weightsStability AICC-BY-SA-4.02.8B29 Sept 20230✓ fits2.1 GB4bit✓ fits3.7 GB8bit
stack-edu-classifier-pythonOpen weightsHugging Faceunclassified124.2M23 Sept 20240✓ fits0.6 GB4bit✓ fits0.6 GB8bit
starencoderOpen weightsbigcodeunclassified0
table-transformer-detectionOpen weightsMicrosoftMIT28.8M14 Oct 20220✓ fits0.5 GB4bit✓ fits0.5 GB8bit
table-transformer-structure-recognitionOpen weightsMicrosoftMIT28.9M14 Oct 20220✓ fits0.5 GB4bit✓ fits0.5 GB8bit
tapex-base-finetuned-wikisqlOpen weightsMicrosoftMIT2 Mar 20220
timesfm-2.5-200m-pytorchOpen weightsGoogleApache-2.0231.3M2 Sept 20250✓ fits0.6 GB4bit✓ fits0.8 GB8bit
tiny-aya-baserestricted-weightsCohereCC-BY-NC-4.03.35B13 Feb 20260✓ fits2.4 GB4bit✓ fits4.3 GB8bit
togethercomputer/LLaMA-2-7B-32Krestricted-weightsTogether AILlama-2-Community7B26 Jul 20230✓ fits4.5 GB4bit✓ fits8.6 GB8bit
togethercomputer/Llama-2-7B-32K-Instructrestricted-weightsTogether AILlama-2-Community7B8 Aug 20230✓ fits4.5 GB4bit✓ fits8.6 GB8bit
unifiedqa-t5-smallOpen weightsAllen Institute for AIunclassified2 Mar 20220
vit-base-patch16-224Open weightsGoogleApache-2.086.6M2 Mar 20220✓ fits0.6 GB4bit✓ fits0.6 GB8bit
vit-base-patch16-224-in21kOpen weightsGoogleApache-2.086.4M2 Mar 20220✓ fits0.6 GB4bit✓ fits0.6 GB8bit
wavlm-base-plusOpen weightsMicrosoftunclassified2 Mar 20220
wavlm-largeOpen weightsMicrosoftunclassified2 Mar 20220
whisper-baseOpen weightsOpenAIApache-2.072.6M26 Sept 20220✓ fits0.6 GB4bit✓ fits0.6 GB8bit
whisper-base.enOpen weightsOpenAIApache-2.072.6M26 Sept 20220✓ fits0.6 GB4bit✓ fits0.6 GB8bit
whisper-largeOpen weightsOpenAIApache-2.01.54B26 Sept 20220✓ fits1.4 GB4bit✓ fits2.3 GB8bit
whisper-large-v2Open weightsOpenAIApache-2.01.54B5 Dec 20220✓ fits1.4 GB4bit✓ fits2.3 GB8bit
whisper-mediumOpen weightsOpenAIApache-2.0763.9M26 Sept 20220✓ fits0.9 GB4bit✓ fits1.4 GB8bit
whisper-medium.enOpen weightsOpenAIApache-2.0763.9M26 Sept 20220✓ fits0.9 GB4bit✓ fits1.4 GB8bit
whisper-smallOpen weightsOpenAIApache-2.0241.7M26 Sept 20220✓ fits0.6 GB4bit✓ fits0.8 GB8bit
whisper-small.enOpen weightsOpenAIApache-2.0241.7M26 Sept 20220✓ fits0.6 GB4bit✓ fits0.8 GB8bit
whisper-tinyOpen weightsOpenAIApache-2.037.8M26 Sept 20220✓ fits0.5 GB4bit✓ fits0.6 GB8bit
whisper-tiny.enOpen weightsOpenAIApache-2.037.8M26 Sept 20220✓ fits0.5 GB4bit✓ fits0.6 GB8bit
wildguardOpen weightsAllen Institute for AIApache-2.07.25B15 Jun 20240✓ fits4.7 GB4bit✓ fits8.8 GB8bit

Best results = rank inside each benchmark's primary comparability group; several dimensions are shown side by side and never combined. Fit columns are estimates (64 GB device at 4-bit, 128 GB at 8-bit, 8K context) — method on /methodology. unclassified = the raw licence label is not yet mapped in the ontology.

MethodUniverse = canonical models whose weights can be downloaded (open-weights, open-source, restricted-weights). hardware_fit values are ESTIMATES (see /methodology); best_results are ranks inside each benchmark's primary comparability group — no composite score. /methodology