Project AURIONAI Use-Cases Research & Innovation
Meta

Meta

Big tech, USA. 184 models in the directory.

big techUSA
0deployments

Identity

Type
big tech
Country label
USA
HQ country
USA
Founding country
USA
Parent
not in the notes
Parent country
not in the notes
Country basis
the research notes
Where this comes from
Convention D1: a dual label, HQ first, when the HQ, founding or parent country differ.

Models by Class

LLM29

Models

ModelClassOpennessParametersReleased
MetaMuse Spark 1.3multimodal LLMproprietary–2 Sep 2026MetaMuse Glimmer 30Bmultimodal LLMopen weights permissive29.6B10 Aug 2026MetaMuse Spark 1.1LLMproprietary–9 Jul 2026MetaMuse Videovideo generationproprietary–7 Jul 2026MetaMuse Imageimage generationproprietary–Jul 2026MetaSapiens2visionopen weights restricted1.46B23 Apr 2026MetaMuse SparkLLMproprietary–8 Apr 2026MetaSAM 3.1 (Object Multiplex)visionopen weights restricted–27 Mar 2026MetaEUPE (Efficient Universal Perception Encoder)visionunknown–23 Mar 2026MetaSAM Audiomusic and audioopen weights restricted–12 Dec 2025MetaOmnilingual ASRspeech recognitionopen source–10 Nov 2025MetaSAM 3visionopen weights restricted848MNov 2025MetaMapAnything3Dopen weights restricted1.23B8 Sep 2025MetaDINOv3visionopen weights restricted6.72B13 Aug 2025MetaDINOv3 ConvNeXt-Bvisionopen weights restricted–13 Aug 2025MetaDINOv3 ConvNeXt-Lvisionopen weights restricted–13 Aug 2025MetaDINOv3 ConvNeXt-Svisionopen weights restricted–13 Aug 2025MetaDINOv3 ConvNeXt-Tvisionopen weights restricted–13 Aug 2025MetaDINOv3 ViT-7Bvisionopen weights restricted6.72B13 Aug 2025MetaDINOv3 ViT-Bvisionopen weights restricted–13 Aug 2025MetaDINOv3 ViT-H+visionopen weights restricted840M13 Aug 2025MetaDINOv3 ViT-Lvisionopen weights restricted–13 Aug 2025MetaDINOv3 ViT-Svisionopen weights restricted21M13 Aug 2025MetaDINOv3 ViT-S+visionopen weights restricted–13 Aug 2025MetaDINOv3 SAT-493M ViT-7Bgeospatialopen weights restricted6.7BAug 2025MetaDINOv3 SAT-493M ViT-L (distilled)geospatialopen weights restricted300MAug 2025MetaMetaCLIP 2 (worldwide, ViT-H)embeddingopen weights restricted1.86B31 Jul 2025MetaV-JEPA 2world modelopen weights permissive1.2B11 Jun 2025MetaV-JEPA 2-ACrobot policyunknown–11 Jun 2025MetaPerception Encoderembeddingopen weights permissive–11 Apr 2025MetaPE-Core-L14-336embeddingopen weights permissive–11 Apr 2025MetaPE-Core-G14-448embeddingopen weights permissive–11 Apr 2025MetaLlama 4LLMopen weights restricted–Apr 2025MetaLlama 4 Maverickmultimodal LLMopen weights restricted400BApr 2025MetaLlama 4 Scoutmultimodal LLMopen weights restricted–Apr 2025MetaLlama Guard 4 12Bguardopen weights restricted12BApr 2025MetaLlama Prompt Guard 2 22Mguardopen weights restricted22MApr 2025MetaLlama Prompt Guard 2 86Mguardopen weights restricted86MApr 2025MetaPrompt-Guard-86Mguardopen weights restricted86MApr 2025MetaeSEN-30M-OAMmaterialsunknown30.2M17 Mar 2025MetaVGGT3Dopen weights restricted1.26B11 Mar 2025MetaAudiobox Aestheticsaudio classificationopen weights permissive103.86M7 Feb 2025MetaEdgeTAMvisionopen source–13 Jan 2025MetaLlama-3.3-70B-InstructLLMunknown–26 Nov 2024MetaMobileLLM 1Bedge LLMunknown–31 Oct 2024MetaMobileLLM 125Medge LLMunknown–30 Oct 2024MetaLlama 3.2LLMunknown–24 Oct 2024MetaLlama 3.2 1Bedge LLMunknown1B24 Oct 2024MetaLlama 3.2 3Bedge LLMunknown–24 Oct 2024MetaLongVUMeta, King Abdullah University of Science and Technologymultimodal LLMopen weights permissive7.67B18 Oct 2024MetaSapiensvisionunknown–22 Aug 2024MetaLlama-3.1-8B-InstructLLMunknown–18 Jul 2024MetaHSTU (Generative Recommenders)recommendationopen source1.5T27 Feb 2024MetaMobileLLMedge LLMunknown–22 Feb 2024MetaMAGNeTmusic and audioopen weights restricted–9 Jan 2024Metaw2v-bert-2.0speech recognitionopen weights permissive600M19 Dec 2023MetaLlama Guardguardunknown–Dec 2023MetaSeamlessM4T v2 largetranslationopen weights restricted2.31B30 Nov 2023New York UniversityDobb-ENew York University, Metarobot policyopen weights permissive21.3M27 Nov 2023MetaNougatOCR and documentopen weights restricted348.74M25 Aug 2023MetaCode Llamacodeopen weights restricted–24 Aug 2023MetaSeamlessM4Ttranslationopen weights restricted–22 Aug 2023MetaDINOv2visionopen weights permissive304.4M17 Jul 2023Metamms-1b-allspeech recognitionunknown–27 May 2023MetaImageBindembeddingopen weights restricted–9 May 2023MetaI-JEPAvisionunknown–19 Jan 2023MetaConvNeXt V2visionopen weights restricted197.96M2 Jan 2023MetaMusicGenmusic and audiounknown–2023Tsinghua UniversityHorNetTsinghua University, Metavisionunknown–28 Jul 2022MetaNLLB-200translationopen weights restricted–8 Jul 2022MetaNLLB-200 distilled 600Mtranslationopen weights restricted–8 Jul 2022MetaNLLB-200 distilled 1.3Btranslationopen weights restricted–8 Jul 2022MetaNLLB-200 3.3Btranslationopen weights restricted–8 Jul 2022MetaOPT 66BLLMunknown–23 Jun 2022MetaOPT 125MLLMunknown–11 May 2022MetaOPT 1.3BLLMunknown–11 May 2022MetaOPT 6.7BLLMunknown–11 May 2022MetaOPTLLMunknown–2 May 2022MetaDeiT IIIvisionunknown–14 Apr 2022MetaViTDetvisionunknown–30 Mar 2022Metadata2vec visionvisionopen weights permissive–7 Feb 2022MetaConvNeXtvisionunknown–10 Jan 2022MetaConvNeXtvisionopen weights permissive–10 Jan 2022MetaDeticvisionunknown–7 Jan 2022MetaMask2Formervisionopen weights restricted216.02M2 Dec 2021MetaMAE (Masked Autoencoder)visionunknown–11 Nov 2021MetaMaskFormervisionopen weights restricted41.78M13 Jul 2021MetaXCiTvisionunknown–17 Jun 2021MetaHuBERTspeech recognitionopen weights permissive–14 Jun 2021Metahubert-base-ls960speech recognitionopen weights permissive–14 Jun 2021Metahubert-large-ll60kspeech recognitionopen weights permissive–14 Jun 2021Metahubert-xlarge-ll60kspeech recognitionopen weights permissive–14 Jun 2021MetaDINO (v1)visionopen weights permissive–29 Apr 2021MetaMViT (Multiscale Vision Transformers)visionunknown–22 Apr 2021MetaLeViTvisionopen source–2 Apr 2021MetaCaiTvisionopen weights permissive356.46M31 Mar 2021Ecole Normale Superieure (Paris)ConViTEcole Normale Superieure (Paris), Metavisionunknown–19 Mar 2021MetaRegNetZvisionunknown–11 Mar 2021MetaBarlow TwinsMeta, New York Universityvisionunknown–4 Mar 2021MetaD2Govisionopen source–12 Feb 2021MetaTimeSformervisionopen weights restricted–9 Feb 2021MetaDeiTvisionunknown–23 Dec 2020MetaDeiT-B distilledvisionopen weights permissive–23 Dec 2020MetaDeiT-Tivisionopen weights permissive5.72M23 Dec 2020Seoul National UniversityInterNet (InterHand2.6M)Seoul National University, Metavisionunknown–21 Aug 2020INRIASwAVINRIA, Metavisionunknown–17 Jun 2020MetaDETRvisionopen weights permissive–26 May 2020MetaDETR ResNet-50visionopen weights permissive41M26 May 2020MetaDETR ResNet-101visionopen weights permissive60.68M26 May 2020MetaDETR ResNet-50 panopticvisionopen weights permissive–26 May 2020MetaResNeStMeta, Amazonvisionunknown–19 Apr 2020MetaX3Dvisionopen source–9 Apr 2020MetaRegNetvisionunknown–30 Mar 2020MetaImVoteNetMeta, Stanford Universityvisionunknown–29 Jan 2020MetaPointRendvisionunknown–17 Dec 2019MetaMoCovisionunknown–13 Nov 2019MetaXLM-RoBERTalanguage encoderopen weights permissive–5 Nov 2019MetaBARTLLMunknown–29 Oct 2019MetaDetectron2 model zoovisionopen source–5 Sep 2019MetaRoBERTalanguage encoderopen weights permissive–26 Jul 2019MetaMesh R-CNN3Dopen weights restricted–6 Jun 2019MetaDLRM (Deep Learning Recommendation Model)recommendationopen source–31 May 2019MetaVoteNetvisionunknown–21 Apr 2019MetaCSNvisionunknown–4 Apr 2019MetaTensorMaskvisionunknown–28 Mar 2019MetaPanoptic FPNvisionunknown–8 Jan 2019MetaLASERembeddingunknown–26 Dec 2018University of Texas at AustinLFB (Long-Term Feature Bank)University of Texas at Austin, Metavisionunknown–12 Dec 2018MetaSlowFastvisionopen source–10 Dec 2018ETH ZurichVideoPose3DETH Zurich, Metavisionunknown–28 Nov 2018MetaResNeXt-101 32x8d WSLvisionunknown–2 May 2018MetaDensePosevisionunknown–1 Feb 2018Carnegie Mellon UniversityNon-local Networks (video)Carnegie Mellon University, Metavisionunknown–21 Nov 2017MetaRetinaNetvisionunknown–7 Aug 2017MetaMask R-CNNvisionunknown–20 Mar 2017MetaKeypoint R-CNNvisionunknown–20 Mar 2017UC San DiegoResNeXtUC San Diego, Metavisionunknown–16 Nov 2016UC San DiegoResNeXt-50 32x4dUC San Diego, Metavisionunknown25M16 Nov 2016Cornell UniversityDenseNetCornell University, Tsinghua University, Metavisionopen weights permissive8.06M25 Aug 2016MetafastText word vectorsembeddingopen weights permissive–15 Jul 2016indicoDCGANindico, Metaimage generationunknown–19 Nov 2015MetaC3DMeta, Dartmouth Collegevisionunknown–2 Dec 2014MetaESM-2 / ESMFoldscientificunknown––MetaLlama 2LLMunknown––MetaLlama 2 70BLLMunknown––MetaLlama-2-7BLLMunknown7B–MetaLlama 3.1 405BLLMunknown––MetaLlama-3.1-405B-InstructLLMunknown––MetaLlama 3.1 70BLLMunknown70B–MetaLlama 3.1 70B InstructLLMunknown70B–MetaLlama 3.1 8BLLMunknown––MetaLlama-3.2-1B-InstructLLMunknown––MetaLlama 3.2 1B quantized (QLoRA)edge LLMunknown––MetaLlama 3.2 1B quantized (SpinQuant)edge LLMunknown––MetaLlama 3.2 3B quantized (QLoRA)edge LLMunknown––MetaLlama 3.2 3B quantized (SpinQuant)edge LLMunknown––MetaLlama3-70B InstructLLMunknown70B–MetaLlama 3 8BLLMunknown8B–MetaMeta-Llama-3-8B-InstructLLMunknown––MetaMMSspeech recognitionunknown––MetaMuse GlimmerLLMunknown––MetaMuse Spark 1.2LLMproprietary––MetaMuse Spark (family)LLMproprietary––MetaProphettime seriesunknown––MetaSAM 1visionunknown––MetaSAM 2visionunknown––MetaSAM 3D Objects3Dunknown––MetaV-JEPAworld modelunknown––Metawav2vec 2.0speech recognitionunknown––Metawav2vec2-large-xlsr-53speech recognitionunknown––MetaSparshrobot policyunknown––MetaRoBERTa baselanguage encoderopen weights permissive124.7M–MetaRoBERTa largelanguage encoderopen weights permissive355.41M–MetaXLM-RoBERTa baselanguage encoderopen weights permissive278.89M–MetaXLM-RoBERTa largelanguage encoderopen weights permissive561.19M–MetaBART largeLLMopen weights permissive––MetaBART large CNNLLMopen weights permissive406.29M–MetaBART large MNLIlanguage encoderopen weights permissive407.34M–MetamBART-50 many-to-manytranslationunknown611.13M–MetaCode Llama 7Bcodeopen weights restricted6.74B–MetaCode Llama 13B Instructcodeopen weights restricted13.02B–MetaCode Llama 34B Pythoncodeopen weights restricted33.74B–MetaCode Llama 70B Instructcodeopen weights restricted68.98B–MetaVoteNetvisionunknown––

Scores

BenchmarkModelScoreWho ran itSetting
Arena Text leaderboardMuse Spark 1.31,494 Eloindependentmax
Arena Vision leaderboardMuse Spark1,294 Eloindependentmuse-spark
Arena Vision leaderboardMuse Spark 1.11,280 Eloindependent
Arena Vision leaderboardMuse Spark 1.21,293 EloindependentxHigh
Arena Vision leaderboardMuse Spark 1.31,286 EloindependentxHigh
Arena Vision leaderboardMuse Spark 1.31,290 Eloindependentmax (muse-spark-1.3-max)
Artificial Analysis Intelligence IndexMuse Spark 1.348 indexindependentmax
Artificial Analysis Intelligence IndexMuse Spark 1.348 indexindependentmax
Artificial Analysis Music ArenaMusicGen881 Eloindependentsecondary capture of AA post; lowest reported
Artificial Analysis Text-to-Image ArenaMuse Image1,116 Eloindependentrank shown as 8-10
Common Voice 22 (CV22-76)Omnilingual ASR6.4 CERvendor-reported7B LLM-ASR, CV22-76, approximate
Epic-Kitchens-100 action anticipationV-JEPA 239.7 R@5vendor-reportedanticipation
FLEURSOmnilingual ASR6.2 CERvendor-reported7B LLM-ASR, FLEURS-102, approximate
IDSV (Indonesian Speech Variability)mms-1b-all29.3 WERindependentIDSV test, all, zero-shot
IDSV (Indonesian Speech Variability)mms-1b-all9.57 CERindependentIDSV test, all, zero-shot
LMArena Text-to-Image ArenaMuse Image1,274 Eloindependent
LMArena Text-to-Video ArenaMuse Video1,456 Eloindependent
LVISSAM 337.2 cgF1vendor-reportedinstance segmentation
LVISSAM 340.6 cgF1vendor-reportedbox
LVISSAM 348.5 APvendor-reportedinstance segmentation
Matbench DiscoveryeSEN-30M-OAM0.17 kSRMEmaintainer-runlower is better
Matbench DiscoveryeSEN-30M-OAM0.888 CPSmaintainer-run
Matbench DiscoveryeSEN-30M-OAM0.925 F1maintainer-run
MCP AtlasMuse Glimmer 30B75.5 %vendor-reported
PerceptionTestV-JEPA 284 %vendor-reported8B LLM-aligned scale
SA-Co/GoldSAM 354.1 cgF1vendor-reportedhuman 72.8
SA-VSAM 330.3 cgF1vendor-reportedtest; human 53.1
SEA-HELMLlama 4 Scout63.43 scoremaintainer-runCheckpoint Llama-4-Scout-17B-16E-Instruct as listed; release view (data last updated 18 Sep 2026); score is the average of 2000 bootstraps
SEA-HELMLlama 4 Scout67.86 scoremaintainer-runCheckpoint Llama-4-Scout-17B-16E-Instruct as listed; release view (data last updated 18 Sep 2026); score is the average of 2000 bootstraps
SEA-HELMLlama-3.1-8B-Instruct37.2 scoremaintainer-runCheckpoint Llama-3.1-8B-Instruct as listed; release view (data last updated 18 Sep 2026); score is the average of 2000 bootstraps
SEA-HELMLlama-3.1-8B-Instruct54.48 scoremaintainer-runCheckpoint Llama-3.1-8B-Instruct as listed; release view (data last updated 18 Sep 2026); score is the average of 2000 bootstraps
SEA-HELMLlama-3.3-70B-Instruct60.44 scoremaintainer-runCheckpoint Llama-3.3-70B-Instruct as listed; release view (data last updated 18 Sep 2026); score is the average of 2000 bootstraps
SEA-HELMLlama-3.3-70B-Instruct71.37 scoremaintainer-runCheckpoint Llama-3.3-70B-Instruct as listed; release view (data last updated 18 Sep 2026); score is the average of 2000 bootstraps
SEA-HELMMuse Glimmer 30B72.13 scoremaintainer-run
SEA-HELMMuse Glimmer 30B75.61 scoremaintainer-run
Something-Something v2V-JEPA 277.3 % top-1vendor-reported
SWE-bench ProMuse Spark 1.161.5 %maintainer-runScale SEAL public, mini-swe-agent, +/-3.10
SWE-bench VerifiedMuse Glimmer 30B76 %vendor-reported
TempCompassV-JEPA 276.9 %vendor-reported8B LLM-aligned scale
Where this comes from
Every recorded score of this organisation's models, with who ran it. Scores from different settings are not directly comparable.

Sources

  1. AI at Meta Blog ai.meta.com https://ai.meta.com/blog/
  2. Muse Spark 1.3 (max) - Intelligence, Performance & Price Analysis | Artificial Analysis artificialanalysis.ai https://artificialanalysis.ai/models/muse-spark-1-3
  3. huggingface.co/meta-models https://huggingface.co/meta-models
  4. huggingface.co/meta-llama https://huggingface.co/meta-llama
  5. GitHub - facebookresearch/sam3: The repository provides code for running inference and finetuning with the Meta Segment Anything Model 3 (SA github.com https://github.com/facebookresearch/sam3
  6. huggingface.co/facebook/dinov3-vit7b16-pretrain-sat493m https://huggingface.co/facebook/dinov3-vit7b16-pretrain-sat493m
  7. Meta Superintelligence Labs - Wikipedia en.wikipedia.org https://en.wikipedia.org/wiki/Meta_Superintelligence_Labs
  8. Introducing quantized Llama models with increased speed and a reduced memory footprint ai.meta.com https://ai.meta.com/blog/meta-llama-quantized-lightweight-models/
  9. GitHub - pytorch/executorch: On-device AI across mobile, embedded and edge for PyTorch github.com https://github.com/pytorch/executorch
  10. Indonesia Tech Leaders Team With NVIDIA and Partners to Launch Nation’s AI blogs.nvidia.com https://blogs.nvidia.com/?p=75648
  11. Training compute of frontier AI models grows by 4-5x per year epoch.ai https://epoch.ai/blog/training-compute-of-frontier-ai-models-grows-by-4-5x-per-year
  12. Over 30 AI models have been trained at the scale of GPT-4 epoch.ai https://epoch.ai/data-insights/models-over-1e25-flop
  13. Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations arxiv.org https://arxiv.org/abs/2312.06674
  14. Introducing the V-JEPA 2 world model and new benchmarks for physical reasoning ai.meta.com https://ai.meta.com/blog/v-jepa-2-world-model-benchmarks/

To cite this page: Project AURION, AI Use-Cases Research & Innovation, "Meta", https://aurion.semai.id/organisations/meta/, records of October 2026.