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.

889 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 (311) · MIT License (125) · Other (unclassified licence) (83) · Creative Commons Attribution-NonCommercial 4.0 (24) · Apple Sample Code / ML Research License (15) · Llama 3.2 Community License (12)

Category · this pageCommercial useRedistributionDerivativesHosting
Open source n=0
Open weights n=5022028220282202822028
restricted-weights n=0

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

Downloadable models

889 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
OASIS-code-embedding-1.5BOpen weightsKwaipilot · embedding, textMIT1.54B11 Mar 20250✓ fits1.4 GB4bit✓ fits2.3 GB8bit
OLMo-1B-hfOpen weightsAllen Institute for AI · textApache-2.01.18B12 Apr 20240✓ fits1.2 GB4bit✓ fits1.9 GB8bit
OLMo-2-0325-32B-DPOOpen weightsAllen Institute for AIunclassified0
OLMo-2-0325-32B-InstructOpen weightsAllen Institute for AI · textApache-2.032.2B12 Mar 20250✓ fits19.0 GB4bit✓ fits37.6 GB8bit
OLMo-2-0425-1BOpen weightsAllen Institute for AI · textApache-2.01.48B17 Apr 20250✓ fits1.4 GB4bit✓ fits2.2 GB8bit
OLMo-2-0425-1B-InstructOpen weightsAllen Institute for AI · textApache-2.01.48B29 Apr 20250✓ fits1.4 GB4bit✓ fits2.2 GB8bit
OLMo-2-0425-1B-RLVR1Open weightsAllen Institute for AIunclassified0
OLMo-2-1124-7BOpen weightsAllen Institute for AIApache-2.07.3B29 Oct 20240✓ fits4.7 GB4bit✓ fits8.9 GB8bit
OLMo-2-1124-7B-DPOOpen weightsAllen Institute for AIunclassified0
OLMo-2-1124-7B-InstructOpen weightsAllen Institute for AI · textApache-2.07.3B18 Dec 20240✓ fits4.7 GB4bit✓ fits8.9 GB8bit
OLMoE-1B-7B-0125-DPOOpen weightsAllen Institute for AIunclassified0
OLMoE-1B-7B-0125-InstructOpen weightsAllen Institute for AI · textApache-2.06.92B27 Jan 20250✓ fits4.5 GB4bit✓ fits8.5 GB8bit
OLMoE-1B-7B-0924Open weightsAllen Institute for AI · textApache-2.06.92B20 Jul 20240✓ fits4.5 GB4bit✓ fits8.5 GB8bit
Olmo-3-1025-7BOpen weightsAllen Institute for AI · textApache-2.07.3B12 Sept 20250✓ fits4.7 GB4bit✓ fits8.9 GB8bit
Olmo-3-7B-Instruct-DPOOpen weightsAllen Institute for AIunclassified0
Olmo-3-7B-Think-DPOOpen weightsAllen Institute for AIunclassified0
Phi-3.5-vision-instructOpen weightsMicrosoft · image, textMIT4.15B16 Aug 20240✓ fits2.9 GB4bit✓ fits5.3 GB8bit
Pythia-Chat-Base-7BOpen weightsTogether AI · textApache-2.07B22 Mar 20230✓ fits4.5 GB4bit✓ fits8.6 GB8bit
QVQ-72B-PreviewOpen weightsQwen Team · image, textunclassified72B25 Dec 20240✓ fits41.9 GB4bit✓ fits83.3 GB8bit
Qianfan-OCROpen weightsBaidu · image, textApache-2.04.74B18 Mar 20260✓ fits3.2 GB4bit✓ fits6.0 GB8bit
Qianfan-VL-3BOpen weightsBaidu · image, textOther3.71B16 Sept 20250✓ fits2.6 GB4bit✓ fits4.8 GB8bit
Qianfan-VL-8BOpen weightsBaidu · image, textOther8.81B16 Sept 20250✓ fits5.6 GB4bit✓ fits10.6 GB8bit
QwQ-Max-PreviewOpen weightsQwen Team · textApache-2.025 Feb 20250
Qwen2-AudioOpen weightsQwen Team · audio, textunclassified7B9 Aug 20240✓ fits4.5 GB4bit✓ fits8.6 GB8bit
Qwen2-VLOpen weightsQwen Team · image, text, videoApache-2.029 Aug 20240
Qwen2.5 7B InstructOpen weightsQwen · textApache-2.07.62B32.8K16 Sept 20241from $0.20 out✓ fits4.9 GB4bit✓ fits9.3 GB8bit
Qwen2.5 VLOpen weightsQwen Team · image, text, videounclassified26 Jan 20250
Qwen2.5 VL 72B InstructOpen weightsQwen · image, textunclassified128K1 Feb 20251from $1 out
Qwen2.5-0.5BOpen weightsQwenunclassified0
Qwen2.5-0.5B-InstructOpen weightsQwen · textApache-2.0494M16 Sept 20240✓ fits0.8 GB4bit✓ fits1.1 GB8bit
Qwen2.5-1.5BOpen weightsQwenunclassified0
Qwen2.5-1.5B-InstructOpen weightsQwen · textApache-2.01.54B17 Sept 20240✓ fits1.4 GB4bit✓ fits2.3 GB8bit
Qwen2.5-1MOpen weightsQwen Team · textunclassified1M27 Jan 20250
Qwen2.5-32BOpen weightsQwenunclassified0
Qwen2.5-3BOpen weightsQwenunclassified0
Qwen2.5-72BOpen weightsQwenunclassified0
Qwen2.5-7BOpen weightsQwenunclassified0
Qwen2.5-MathOpen weightsQwen Team · image, textunclassified4.1K19 Sept 20240
Qwen2.5-Math-7BOpen weightsQwenunclassified0
Qwen2.5-OmniOpen weightsQwen Team · audio, image, text, videounclassified7B27 Mar 20250✓ fits4.5 GB4bit✓ fits8.6 GB8bit
Qwen2.5-VL-32B-InstructOpen weightsQwen Team · image, textApache-2.032B24 Mar 20250✓ fits18.9 GB4bit✓ fits37.3 GB8bit
Qwen2.5-VL-7B-InstructOpen weightsQwen · image, textApache-2.08.29B26 Jan 20250✓ fits5.3 GB4bit✓ fits10.0 GB8bit
Qwen3 235B A22BOpen weightsQwen · textApache-2.0235B22B active131.1K29 Apr 20251from $1.82 out✗ too large135.6 GB4bit✗ too large270.8 GB8bit
Qwen3 235B A22B Thinking 2507Open weightsQwen · textunclassified131.1K25 Jul 20251from $2.3 out
Qwen3 30B A3BOpen weightsQwen · textunclassified131.1K28 Apr 20251from $0.50 out
Qwen3 30B A3B Instruct 2507Open weightsQwen · textunclassified262.1K29 Jul 20251from $0.30 out
Qwen3 30B A3B Thinking 2507Open weightsQwen · textunclassified81.9K28 Aug 20251from $2.4 out
Qwen3 Coder 480B A35BOpen weightsQwen · textunclassified262.1K23 Jul 20251from $1 out
Qwen3 Embedding SeriesOpen weightsQwen Team · embedding, score, textApache-2.032.8K5 Jun 20250
Qwen3 Next 80B A3B ThinkingOpen weightsQwen · textunclassified262.1K11 Sept 20251from $1.2 out

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