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

Open source

1

Open weights

869

Restricted weights

80

Downloadable totalUniverse = 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.

950

New in 30 daysDownloadable models with a release date inside the last 30 days (release_date claim; models without one are not counted).

269

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=5015035150351503515035
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

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
DeepSeek-V4.1-FlashOpen weightsDeepSeek · image, text—unclassified—1M10 Sept 2026livebench-agentic-coding#1livebench-2#5livebench-data-analysis#122from $0.60 out——
Gemma 3 27BOpen weightsGoogle · image, text—unclassified—131.1K12 Mar 2025aider-polyglot-well-formed#1aider-polyglot#60scicode#1201from $0.45 out——
Z.ai GLM 5.2Open weightsZ.ai (Zhipu AI) · textMIT753.3B1M16 Jun 2026tau2-bench#1terminal-bench#12livebench-coding#162from $2 out✗ too large433.7 GB4bit✗ too large866.8 GB8bit
MiniMax M2.5Open weightsMiniMax · textOther228.7B204.8K12 Feb 2026swe-bench-verified#2swe-bench-multilingual#6tau2-bench#221from $1.2 out✗ too large132.0 GB4bit✗ too large263.5 GB8bit
GLM 4.7 FlashOpen weightsZ.ai (Zhipu AI) · textMIT31.2B200K19 Jan 2026tau2-bench#3ifbench#100terminal-bench#1141from $0.40 out✓ fits18.4 GB4bit✓ fits36.4 GB8bit
MiniMax-M3Open weightsMiniMax · image, text, videoOther427B1.05M2 Jun 2026ifbench#3gpqa-diamond#14mmmu-pro#253from $1.2 out✗ too large246.1 GB4bit✗ too large491.6 GB8bit
DeepSeek V4 Flash Vision ExpOpen weightsDeepSeek · image, textMIT304.6B1.05M31 Aug 2026livebench-agentic-coding#4livebench-data-analysis#8livebench-2#212from $0.33 out✗ too large175.7 GB4bit✗ too large350.9 GB8bit
GLM 5Open weightsZ.ai (Zhipu AI) · textMIT753.9B204.8K11 Feb 2026swe-bench-multilingual#4swe-bench-verified#6tau2-bench#81from $1.92 out✗ too large434.0 GB4bit✗ too large867.4 GB8bit
Nemotron 3 UltraOpen weightsNVIDIA · text—unclassified—262.1K4 Jun 2026ifbench#4livebench-if#14livebench-mathematics#311from $3.13 out——
Step 3.7 FlashOpen weightsStepFun · image, text, video—unclassified—262.1K28 May 2026tau2-bench#4mmmu-pro#44terminal-bench#521from $1.15 out——
Kimi K3Open weightsMoonshot AI · image, text, videoOther2.78T1.05M13 Jun 2026scicode#5livebench-reasoning#5livebench-agentic-coding#83from $11.55 out✗ too large1599.0 GB4bit✗ too large3197.4 GB8bit
Qwen3.8 FlashOpen weightsQwen · image, text, video—unclassified—1M26 Aug 2026livebench-if#5livebench-agentic-coding#10artificial-analysis-intelligence-index#172from $0.47 out——
GLM 5.3Open weightsZ.ai (Zhipu AI) · textOther753.3B1.31M25 Aug 2026scicode#6artificial-analysis-intelligence-index#7livebench-agentic-coding#133from $2.2 out✗ too large433.7 GB4bit✗ too large866.8 GB8bit
Kimi K2.5Open weightsMoonshot AI · image, textOther1.03T262.1K1 Jan 2026swe-bench-multilingual#7swe-bench-verified#10tau2-bench#141from $2.25 out✗ too large591.0 GB4bit✗ too large1181.4 GB8bit
nemotron-cascade-2-30b-a3bOpen weightsNVIDIA—unclassified—1M—ifbench#7terminal-bench#116tau2-bench#1390——
DeepSeek-V4-Pro-0813Open weightsDeepSeek · text—unclassified—1M12 Aug 2026livebench-mathematics#8livebench-data-analysis#13livebench-2#173from $1.73 out——
MiMo-V2.5-ProOpen weightsXiaomi · text—unclassified—1.05M22 Apr 2026ifbench#8terminal-bench#27tau2-bench#311from $0.87 out——
DeepSeek V4 Flash (0731)Open weightsDeepSeek · textMIT304.2B1.31M31 Jul 2026livebench-data-analysis#9livebench-reasoning#25livebench-language#303from $0.08 out✗ too large175.4 GB4bit✗ too large350.3 GB8bit
GLM 5.1Open weightsZ.ai (Zhipu AI) · textMIT753.9B204.8K3 Apr 2026tau2-bench#9ifbench#20terminal-bench#271from $3.04 out✗ too large434.0 GB4bit✗ too large867.4 GB8bit
Qwen3.8 2.4T A95BOpen weightsQwen · text—unclassified—1.05M12 Aug 2026gpqa-diamond#10artificial-analysis-intelligence-index#16scicode#202from $6 out——
deepseek-v4-flash-0420Open weightsDeepSeek—unclassified—1M—ifbench#10tau2-bench#24terminal-bench#520——
DeepSeek V3 0324Open weightsDeepSeek · textMIT684.5B163.8K24 Mar 2025aider-polyglot-well-formed#11aider-polyglot#23scicode#841from $1 out✗ too large394.1 GB4bit✗ too large787.7 GB8bit
DeepSeek V3.2Open weightsDeepSeek · textMIT685.4B163.8K1 Dec 2025swe-bench-verified#11swe-bench-multilingual#12terminal-bench#521from $0.40 out✗ too large394.6 GB4bit✗ too large788.7 GB8bit
Qwen2.5 Coder 32B InstructOpen weightsQwen · text—unclassified—32.8K11 Nov 2024aider-polyglot-well-formed#11swe-bench-verified#42aider-polyglot#531from $1 out——
Qwen3.5 397B A17BOpen weightsQwen · image, text, video—unclassified—262.1K16 Feb 2026ifbench#11tau2-bench#18mmmu-pro#322from $3.5 out——
Qwen3.8 27BOpen weightsQwen · image, text, videoApache-2.027.8B1M5 Aug 2026livebench-agentic-coding#11livebench-if#16livebench-data-analysis#242from $2.55 out✓ fits16.5 GB4bit✓ fits32.5 GB8bit
GLM 5.3 FlashOpen weightsZ.ai (Zhipu AI) · image, text, videoMIT321.3B1.31M25 Aug 2026artificial-analysis-intelligence-index#12livebench-coding#19livebench-agentic-coding#213from $0.25 out✗ too large185.3 GB4bit✗ too large370.0 GB8bit
R1 0528Open weightsDeepSeek · text—unclassified—163.8K28 May 2025aider-polyglot#12aider-polyglot-well-formed#391from $2.15 out——
deepseek-v4-pro-0424Open weightsDeepSeek—unclassified—1M—tau2-bench#13ifbench#18terminal-bench#190——
GLM 4.7Open weightsZ.ai (Zhipu AI) · textMIT358.3B204.8K22 Dec 2025tau2-bench#14ifbench#72terminal-bench#751from $1.75 out✗ too large206.6 GB4bit✗ too large412.6 GB8bit
Kimi K2.6Open weightsMoonshot AI · image, textOther1.03T262.1K14 Apr 2026tau2-bench#14mmmu-pro#20ifbench#212from $4 out✗ too large591.0 GB4bit✗ too large1181.4 GB8bit
Llama 4 MaverickOpen weightsMeta AI · image, text—unclassified—1.05M5 Apr 2025aider-polyglot-well-formed#17aider-polyglot#54mmmu-pro#1001from $0.696 out——
deepseek-v4-flash-0420-highOpen weightsDeepSeek—unclassified—1M—tau2-bench#18ifbench#37terminal-bench#400——
Kimi K2 ThinkingOpen weightsMoonshot AI · textOther1.03T262.1K4 Nov 2025swe-bench-verified#19tau2-bench#41ifbench#711from $2.5 out✗ too large590.7 GB4bit✗ too large1180.9 GB8bit
MiniMax M2Open weightsMiniMax · textOther228.7B204.8K22 Oct 2025swe-bench-verified#20ifbench#46tau2-bench#721from $1.02 out✗ too large132.0 GB4bit✗ too large263.5 GB8bit
Kimi K2 0711Open weightsMoonshot AI · textOther1.03T131.1K11 Jul 2025aider-polyglot#21swe-bench-verified#32aider-polyglot-well-formed#451from $2.3 out✗ too large590.7 GB4bit✗ too large1180.9 GB8bit
Qwen3.6 35B A3BOpen weightsQwen · image, text, videoApache-2.036B3B active262.1K15 Apr 2026tau2-bench#22mmmu-pro#45terminal-bench#571from $0.90 out✓ fits21.2 GB4bit✓ fits41.8 GB8bit
R1Open weightsDeepSeek · textMIT684.5B64K20 Jan 2025aider-polyglot#22aider-polyglot-well-formed#321from $2.5 out✗ too large394.1 GB4bit✗ too large787.7 GB8bit
Kimi K2.7 CodeOpen weightsMoonshot AI · image, textOther1.03T262.1K11 Jun 2026terminal-bench#23livebench-language#34livebench-reasoning#362from $3.5 out✗ too large591.0 GB4bit✗ too large1181.4 GB8bit
Ling 3.0 Flash VLOpen weightsinclusionAI · image, text, video—unclassified—131.1K10 Sept 2026mmmu-pro#23scicode#63artificial-analysis-intelligence-index#811from $0.18 out——
mimo-v2-flashOpen weightsXiaomi—unclassified—256K—tau2-bench#24ifbench#90terminal-bench#920——
GLM 4.6Open weightsZ.ai (Zhipu AI) · text—unclassified—204.8K30 Sept 2025swe-bench-verified#25terminal-bench#87tau2-bench#1071from $1.75 out——
MiniMax M2.7Open weightsMiniMax · textOther228.7B204.8K9 Apr 2026ifbench#25scicode#37terminal-bench#392from $1.2 out✗ too large132.0 GB4bit✗ too large263.5 GB8bit
Qwen3.5-122B-A10BOpen weightsQwen · image, text, video—unclassified—262.1K25 Feb 2026ifbench#25tau2-bench#37mmmu-pro#471from $2.08 out——
qwen3-coder-480b-a35b-instructOpen weightsAlibaba Group · text—unclassified480B35B active256K22 Jul 2025swe-bench-verified#25terminal-bench#123tau2-bench#1560✗ too large276.5 GB4bit✗ too large552.5 GB8bit
InklingOpen weightsThinking Machines · audio, image, text—unclassified—1.05M17 Jul 2026livebench-if#26livebench-mathematics#32livebench-agentic-coding#342from $4.05 out——
GLM 4.5Open weightsZ.ai (Zhipu AI) · textMIT358.3B131.1K20 Jul 2025swe-bench-verified#27terminal-bench#114gpqa-diamond#1541from $2.2 out✗ too large206.6 GB4bit✗ too large412.6 GB8bit
Gemma 4 31BOpen weightsGoogle · image, text, videoApache-2.031.3B262.1K11 Mar 2026ifbench#27terminal-bench#49scicode#552from $0.34 out✓ fits18.5 GB4bit✓ fits36.5 GB8bit
Qwen3.5-27BOpen weightsQwen · image, text, video—unclassified—262.1K25 Feb 2026ifbench#27tau2-bench#35mmmu-pro#451from $1.56 out——
Devstral 2Open weightsMistral AI · text—unclassified—262.1K9 Dec 2025swe-bench-verified#28scicode#104terminal-bench#1231from $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