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=3919020190201902019020
restricted-weights n=11650110011001100

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
qwen-chat-72bOpen weightsAlibaba Groupunclassified33.8Kartificial-analysis-intelligence-index#4220
deepseek-llm-67b-chatOpen weightsDeepSeekunclassified4.1Kartificial-analysis-intelligence-index#4300
llama-65bOpen weightsMeta AIunclassified2.05Kartificial-analysis-intelligence-index#4520
qwen-chat-14bOpen weightsAlibaba Groupunclassified8.19Kartificial-analysis-intelligence-index#4520
3b-de-ft-research_releaseOpen weightsCanopy LabsApache-2.03.3B2 Apr 20250✓ fits2.4 GB4bit✓ fits4.3 GB8bit
3b-de-pretrain-research_releaserestricted-weightsCanopy LabsLlama-3.2-Community3.78B23 Mar 20250✓ fits2.7 GB4bit✓ fits4.8 GB8bit
3b-es_it-ft-research_releaseOpen weightsCanopy LabsApache-2.03.3B4 Apr 20250✓ fits2.4 GB4bit✓ fits4.3 GB8bit
3b-es_it-pretrain-research_releaserestricted-weightsCanopy LabsLlama-3.2-Community3.78B27 Mar 20250✓ fits2.7 GB4bit✓ fits4.8 GB8bit
3b-fr-ft-research_releaseOpen weightsCanopy LabsApache-2.03.3B3 Apr 20250✓ fits2.4 GB4bit✓ fits4.3 GB8bit
3b-fr-pretrain-research_releaserestricted-weightsCanopy LabsLlama-3.2-Community3.78B2 Apr 20250✓ fits2.7 GB4bit✓ fits4.8 GB8bit
3b-hi-ft-research_releaseOpen weightsCanopy LabsApache-2.03.3B8 Apr 20250✓ fits2.4 GB4bit✓ fits4.3 GB8bit
3b-hi-pretrain-research_releaserestricted-weightsCanopy LabsLlama-3.2-Community3.78B8 Apr 20250✓ fits2.7 GB4bit✓ fits4.8 GB8bit
3b-ko-ft-research_releaseOpen weightsCanopy LabsApache-2.03.3B6 Apr 20250✓ fits2.4 GB4bit✓ fits4.3 GB8bit
3b-ko-pretrain-research_releaseOpen weightsCanopy LabsApache-2.03.78B2 Apr 20250✓ fits2.7 GB4bit✓ fits4.8 GB8bit
3b-zh-ft-research_releaseOpen weightsCanopy LabsApache-2.03.3B6 Apr 20250✓ fits2.4 GB4bit✓ fits4.3 GB8bit
3b-zh-pretrain-research_releaserestricted-weightsCanopy LabsLlama-3.2-Community3.78B3 Apr 20250✓ fits2.7 GB4bit✓ fits4.8 GB8bit
AI21-Jamba-Large-1.5Open weightsAI21 LabsOther398.6B19 Aug 20240✗ too large229.7 GB4bit✗ too large458.8 GB8bit
AI21-Jamba-Large-1.6Open weightsAI21 LabsOther398.6B27 Feb 20250✗ too large229.7 GB4bit✗ too large458.8 GB8bit
AI21-Jamba-Large-1.7Open weightsAI21 LabsOther398.6B2 Jul 20250✗ too large229.7 GB4bit✗ too large458.8 GB8bit
AI21-Jamba-Mini-1.5Open weightsAI21 LabsOther51.6B19 Aug 20240✓ fits30.2 GB4bit✓ fits59.8 GB8bit
AI21-Jamba-Mini-1.6Open weightsAI21 LabsOther51.6B27 Feb 20250✓ fits30.2 GB4bit✓ fits59.8 GB8bit
AI21-Jamba-Mini-1.7Open weightsAI21 LabsOther51.6B1 Jul 20250✓ fits30.2 GB4bit✓ fits59.8 GB8bit
AI21-Jamba-Reasoning-3BOpen weightsAI21 LabsApache-2.03.2B5 Oct 20250✓ fits2.3 GB4bit✓ fits4.2 GB8bit
AI21-Jamba2-3BOpen weightsAI21 LabsApache-2.03.03B6 Jan 20260✓ fits2.2 GB4bit✓ fits4.0 GB8bit
AI21-Jamba2-MiniOpen weightsAI21 LabsApache-2.051.6B6 Jan 20260✓ fits30.2 GB4bit✓ fits59.8 GB8bit
AREX-TurboOpen weightsBAAIunclassified0
Aurora-Spec-Minimax-M2.5Open weightsTogether AIApache-2.0857.9M19 Feb 20260✓ fits1.0 GB4bit✓ fits1.5 GB8bit
BAGEL-7B-MoTOpen weightsByteDanceApache-2.014.7B19 May 20250✓ fits8.9 GB4bit✓ fits17.4 GB8bit
BFS-Prover-V2-7BOpen weightsByteDanceApache-2.07.62B6 Oct 20250✓ fits4.9 GB4bit✓ fits9.3 GB8bit
BigBang-v1Open weightsendless-frontierunclassified0
BioMedLMrestricted-weightsStanford CRFMBigScience-BLOOM-RAIL-1.014 Dec 20220
BiomedNLP-BiomedBERT-base-uncased-abstractOpen weightsMicrosoftMIT2 Mar 20220
Cohere-embed-english-v3.0Open weightsCohereunclassified2 Nov 20230
Cohere-embed-multilingual-v3.0Open weightsCohereunclassified2 Nov 20230
Cosmos-Reason2-2BOpen weightsNVIDIAOther2.44B12 Dec 20250✓ fits1.9 GB4bit✓ fits3.3 GB8bit
Cosmos3-EdgeOpen weightsNVIDIAOther3.86B1 Jul 20260✓ fits2.7 GB4bit✓ fits4.9 GB8bit
Cydonia 24B V4.1Open weightsTheDrummer · textunclassified131.1K27 Sept 20251from $0.50 out
DFN2B-CLIP-ViT-B-16restricted-weightsAppleApple-AMLR31 Oct 20230
DFN2B-CLIP-ViT-L-14restricted-weightsAppleApple-AMLR30 Oct 20230
DFN2B-CLIP-ViT-L-14-39Brestricted-weightsAppleApple-AMLR39B8 Jul 20240✓ fits22.9 GB4bit✓ fits45.4 GB8bit
DFN5B-CLIP-ViT-H-14restricted-weightsAppleApple-AMLR30 Oct 20230
DFN5B-CLIP-ViT-H-14-378restricted-weightsAppleApple-AMLR30 Oct 20230
DeepSeek V4 Flash 0423Open weightsDeepSeek · textMIT290.9B1.05M22 Apr 20261from $0.134 out✗ too large167.8 GB4bit✗ too large335.1 GB8bit
DeepSeek V4 Pro 0423Open weightsDeepSeek · textMIT1.6T1.05M22 Apr 20261from $1.64 out✗ too large919.8 GB4bit✗ too large1839.2 GB8bit
DeepSeek-Coder-V2-Lite-InstructOpen weightsDeepSeekOther15.7B14 Jun 20240✓ fits9.5 GB4bit✓ fits18.6 GB8bit
DeepSeek-OCROpen weightsDeepSeekMIT3.34B17 Oct 20250✓ fits2.4 GB4bit✓ fits4.3 GB8bit
DeepSeek-OCR-2Open weightsDeepSeekApache-2.03.39B27 Jan 20260✓ fits2.4 GB4bit✓ fits4.4 GB8bit
DeepSeek-R1-0528Open weightsDeepSeekunclassified0
DeepSeek-R1-Distill-Qwen-7BOpen weightsDeepSeekMIT7.62B20 Jan 20250✓ fits4.9 GB4bit✓ fits9.3 GB8bit
DeepSeek-V2-LiteOpen weightsDeepSeekOther15.7B15 May 20240✓ fits9.5 GB4bit✓ fits18.6 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