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.
79 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.
- Open source1 0%
- Open weights870 92%
- Restricted weights79 8%
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 page | Commercial use | Redistribution | Derivatives | Hosting |
|---|---|---|---|---|
| Open source n=0 | — | — | — | — |
| Open weights n=0 | — | — | — | — |
| restricted-weights n=50 | 32180 | 4802 | 4802 | 1490 |
Counts read: allowed · restricted · unknown / unclassified — over the 50 models on this page, from each licence's stated terms (ontology).
Downloadable models
79 models · sorted by best benchmark rank
Permissions, in order:commercial use· redistribution· derivatives· hosting allowed restricted unknown
| Model | Licence · permissions | Params | Context | Release | Best results | Providers | Fit 64 GB @4bit · 128 GB @8bit Estimated | |
|---|---|---|---|---|---|---|---|---|
| Gemma 3 4Brestricted-weightsGoogle · image, text | Gemma-Terms | 4.3B | 131.1K | 20 Feb 2025 | mmmu-pro#151terminal-bench#265ifbench#287 | 2from $0.10 out | ✓ fits3.0 GB4bit✓ fits5.5 GB8bit | |
| Llama-3.1-405Brestricted-weightsMeta AI | Llama-3.1-Community | 405.9B | 128K | 16 Jul 2024 | terminal-bench#186ifbench#209tau2-bench#257 | 0 | ✗ too large233.9 GB4bit✗ too large467.2 GB8bit | |
| Hermes-4-70Brestricted-weightsNous Research | Llama-3-Community | 70.5B | 128K | 18 Aug 2025 | terminal-bench#210humanitys-last-exam#221gpqa-diamond#228 | 0 | ✓ fits41.1 GB4bit✓ fits81.6 GB8bit | |
| Llama-3.2-3Brestricted-weightsMeta AI | Llama-3.2-Community | 3.21B | 128K | 18 Sept 2024 | tau2-bench#245humanitys-last-exam#294ifbench#299 | 0 | ✓ fits2.4 GB4bit✓ fits4.2 GB8bit | |
| Llama 3.1 8Brestricted-weightsMeta AI | Llama-3.1-Community | 8.03B | 128K | 14 Jul 2024 | terminal-bench#265tau2-bench#265ifbench#286 | 0 | ✓ fits5.1 GB4bit✓ fits9.7 GB8bit | |
| Llama-3.2-1Brestricted-weightsMeta AI | Llama-3.2-Community | 1.24B | 128K | 18 Sept 2024 | terminal-bench#279humanitys-last-exam#290tau2-bench#302 | 0 | ✓ fits1.2 GB4bit✓ fits1.9 GB8bit | |
| tiny-aya-globalrestricted-weightsCohere · text | CC-BY-NC-4.0 | 3.35B | 8K | 13 Feb 2026 | terminal-bench#279humanitys-last-exam#300tau2-bench#302 | 0 | ✓ fits2.4 GB4bit✓ fits4.3 GB8bit | |
| 3b-de-pretrain-research_releaserestricted-weightsCanopy Labs | Llama-3.2-Community | 3.78B | — | 23 Mar 2025 | — | 0 | ✓ fits2.7 GB4bit✓ fits4.8 GB8bit | |
| 3b-es_it-pretrain-research_releaserestricted-weightsCanopy Labs | Llama-3.2-Community | 3.78B | — | 27 Mar 2025 | — | 0 | ✓ fits2.7 GB4bit✓ fits4.8 GB8bit | |
| 3b-fr-pretrain-research_releaserestricted-weightsCanopy Labs | Llama-3.2-Community | 3.78B | — | 2 Apr 2025 | — | 0 | ✓ fits2.7 GB4bit✓ fits4.8 GB8bit | |
| 3b-hi-pretrain-research_releaserestricted-weightsCanopy Labs | Llama-3.2-Community | 3.78B | — | 8 Apr 2025 | — | 0 | ✓ fits2.7 GB4bit✓ fits4.8 GB8bit | |
| 3b-zh-pretrain-research_releaserestricted-weightsCanopy Labs | Llama-3.2-Community | 3.78B | — | 3 Apr 2025 | — | 0 | ✓ fits2.7 GB4bit✓ fits4.8 GB8bit | |
| BioMedLMrestricted-weightsStanford CRFM | BigScience-BLOOM-RAIL-1.0 | — | — | 14 Dec 2022 | — | 0 | —— | |
| DFN2B-CLIP-ViT-B-16restricted-weightsApple | Apple-AMLR | — | — | 31 Oct 2023 | — | 0 | —— | |
| DFN2B-CLIP-ViT-L-14restricted-weightsApple | Apple-AMLR | — | — | 30 Oct 2023 | — | 0 | —— | |
| DFN2B-CLIP-ViT-L-14-39Brestricted-weightsApple | Apple-AMLR | 39B | — | 8 Jul 2024 | — | 0 | ✓ fits22.9 GB4bit✓ fits45.4 GB8bit | |
| DFN5B-CLIP-ViT-H-14restricted-weightsApple | Apple-AMLR | — | — | 30 Oct 2023 | — | 0 | —— | |
| DFN5B-CLIP-ViT-H-14-378restricted-weightsApple | Apple-AMLR | — | — | 30 Oct 2023 | — | 0 | —— | |
| DepthProrestricted-weightsApple | Apple-AMLR | — | — | 3 Oct 2024 | — | 0 | —— | |
| DepthPro-hfrestricted-weightsApple | Apple-AMLR | 952M | — | 27 Nov 2024 | — | 0 | ✓ fits1.1 GB4bit✓ fits1.6 GB8bit | |
| FastVLM-0.5Brestricted-weightsApple | Apple-AMLR | 758.8M | — | 25 Aug 2025 | — | 0 | ✓ fits0.9 GB4bit✓ fits1.4 GB8bit | |
| FastVLM-1.5Brestricted-weightsApple | Apple-AMLR | 1.91B | — | 25 Aug 2025 | — | 0 | ✓ fits1.6 GB4bit✓ fits2.7 GB8bit | |
| FireLLaVA-13brestricted-weightsFireworks AI | Llama-2-Community | 13.3B | — | 5 Jan 2024 | — | 0 | ✓ fits8.2 GB4bit✓ fits15.8 GB8bit | |
| Hermes-3-Llama-3.1-8Brestricted-weightsNous Research | Llama-3-Community | 8.03B | — | 28 Jul 2024 | — | 0 | ✓ fits5.1 GB4bit✓ fits9.7 GB8bit | |
| Llama 3.1 70B Instructrestricted-weightsMeta AI · text | Llama-3.1-Community | 70.5B | 131.1K | 23 Jul 2024 | — | 1from $0.72 out | ✓ fits41.1 GB4bit✓ fits81.6 GB8bit | |
| Llama 3.1 8B Instructrestricted-weightsMeta AI · text | Llama-3.1-Community | 8.03B | 131.1K | 23 Jul 2024 | — | 1from $0.08 out | ✓ fits5.1 GB4bit✓ fits9.7 GB8bit | |
| Llama 3.2 1B Instructrestricted-weightsMeta AI · text | Llama-3.2-Community | 1.24B | 60K | 18 Sept 2024 | — | 1from $0.201 out | ✓ fits1.2 GB4bit✓ fits1.9 GB8bit | |
| Llama 3.2 3B Instructrestricted-weightsMeta AI · text | Llama-3.2-Community | 3.21B | 131.1K | 18 Sept 2024 | — | 1from $0.33 out | ✓ fits2.4 GB4bit✓ fits4.2 GB8bit | |
| Llama 3.3 70B Instructrestricted-weightsMeta AI · text | Llama-3.3-Community | 70.5B | 131.1K | 26 Nov 2024 | — | 1from $0.32 out | ✓ fits41.1 GB4bit✓ fits81.6 GB8bit | |
| Llama-2-13b-chat-hfrestricted-weightsMeta AI | Llama-2-Community | 13B | — | 13 Jul 2023 | — | 0 | ✓ fits8.0 GB4bit✓ fits15.5 GB8bit | |
| Llama-2-7b-chat-hfrestricted-weightsMeta AI | Llama-2-Community | 6.74B | — | 13 Jul 2023 | — | 0 | ✓ fits4.4 GB4bit✓ fits8.3 GB8bit | |
| Llama-2-7b-hfrestricted-weightsMeta AI | Llama-2-Community | 6.74B | — | 13 Jul 2023 | — | 0 | ✓ fits4.4 GB4bit✓ fits8.3 GB8bit | |
| Llama-3.2-11B-Vision-Instructrestricted-weightsMeta AI | Llama-3.2-Community | 10.7B | — | 18 Sept 2024 | — | 0 | ✓ fits6.6 GB4bit✓ fits12.8 GB8bit | |
| Llama-3.2-90B-Vision-Instructrestricted-weightsMeta AI | Llama-3.2-Community | 88.6B | — | 19 Sept 2024 | — | 0 | ✓ fits51.4 GB4bit✓ fits102.4 GB8bit | |
| Meta-Llama-3-70Brestricted-weightsMeta AI | Llama-3-Community | 70.5B | — | 17 Apr 2024 | — | 0 | ✓ fits41.1 GB4bit✓ fits81.6 GB8bit | |
| Meta-Llama-3-8Brestricted-weightsMeta AI | Llama-3-Community | 8.03B | — | 17 Apr 2024 | — | 0 | ✓ fits5.1 GB4bit✓ fits9.7 GB8bit | |
| Meta-Llama-3-8B-Instructrestricted-weightsMeta AI | Llama-3-Community | 8.03B | — | 17 Apr 2024 | — | 0 | ✓ fits5.1 GB4bit✓ fits9.7 GB8bit | |
| MobileCLIP-S1-OpenCLIPrestricted-weightsApple | Apple-AMLR | — | — | 7 Jun 2024 | — | 0 | —— | |
| MobileCLIP-S2-OpenCLIPrestricted-weightsApple | Apple-AMLR | — | — | 7 Jun 2024 | — | 0 | —— | |
| NousResearch/Llama-3.2-1Brestricted-weightsNous Research | Llama-3.2-Community | 1.24B | — | 27 Sept 2024 | — | 0 | ✓ fits1.2 GB4bit✓ fits1.9 GB8bit | |
| NousResearch/Meta-Llama-3.1-70B-Instructrestricted-weightsNous Research | Llama-3.1-Community | 70.5B | — | 24 Jul 2024 | — | 0 | ✓ fits41.1 GB4bit✓ fits81.6 GB8bit | |
| NousResearch/Meta-Llama-3.1-8Brestricted-weightsNous Research | Llama-3.1-Community | 8.03B | — | 24 Jul 2024 | — | 0 | ✓ fits5.1 GB4bit✓ fits9.7 GB8bit | |
| NousResearch/Meta-Llama-3.1-8B-Instructrestricted-weightsNous Research | Llama-3.1-Community | 8.03B | — | 24 Jul 2024 | — | 0 | ✓ fits5.1 GB4bit✓ fits9.7 GB8bit | |
| OpenELM-1_1B-Instructrestricted-weightsApple | Apple-AMLR | 1.08B | — | 12 Apr 2024 | — | 0 | ✓ fits1.1 GB4bit✓ fits1.7 GB8bit | |
| OpenELM-270Mrestricted-weightsApple | Apple-AMLR | 271.5M | — | 12 Apr 2024 | — | 0 | ✓ fits0.7 GB4bit✓ fits0.8 GB8bit | |
| Prompt-Guard-86Mrestricted-weightsMeta AI | Llama-3.1-Community | 278.8M | — | 21 Jul 2024 | — | 0 | ✓ fits0.7 GB4bit✓ fits0.8 GB8bit | |
| Qwen2.5restricted-weightsQwen Team · text | Research-Only | 72.7B70B active | 131.1K | 19 Sept 2024 | — | 0 | ✓ fits42.3 GB4bit✓ fits84.1 GB8bit | |
| Qwen2.5-Coderrestricted-weightsQwen Team · text | Research-Only | 32.5B31B active | 128K | 12 Nov 2024 | — | 0 | ✓ fits19.2 GB4bit✓ fits37.9 GB8bit | |
| SFR-Embedding-2_Rrestricted-weightsSalesforce | CC-BY-NC-4.0 | 7.11B | — | 14 Jun 2024 | — | 0 | ✓ fits4.6 GB4bit✓ fits8.7 GB8bit | |
| SFR-Embedding-Code-400M_Rrestricted-weightsSalesforce | CC-BY-NC-4.0 | 434.1M | — | 16 Jan 2025 | — | 0 | ✓ fits0.8 GB4bit✓ fits1.0 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