← All lab files

Meta research file

#2 on the current provider board.

Muse Spark 1.1 scores 1493.1 according to LMArena's text style-control snapshot published July 16, 2026.

LMArena source rank#2
LMArena rating1493.1
Reported interval1484.9–1501.3
Battles5,732

Model records from the Epoch AI catalog

Meta models in the reviewed record.

Release date, domains, tasks, access, and weight availability come from the Epoch AI catalog. These fields describe models but do not rank them.

  1. Muse Spark

    Multimodal, Language, VisionAPI accessClosed weightsRead profile →

  2. Meta's Generative Ads Model (GEM)

    RecommendationUnreleasedClosed weightsOpen record →

  3. Llama 4 Behemoth (preview)

    Multimodal, Language, VisionUnreleasedClosed weightsOpen record →

  4. Llama 4 Maverick

    Multimodal, Language, VisionOpen weights (restricted use)Open weightsRead profile →

  5. Llama 4 Scout

    Multimodal, Language, VisionOpen weights (restricted use)Open weightsRead profile →

  6. Llama 3.3 70B

    LanguageOpen weights (restricted use)Open weightsOpen record →

  7. Movie Gen Video

    Video, VisionUnreleasedClosed weightsOpen record →

  8. Llama 3.2 11B

    Multimodal, Vision, LanguageOpen weights (restricted use)Open weightsOpen record →

  9. Llama 3.1-405B

    LanguageOpen weights (restricted use)Open weightsOpen record →

  10. Llama 3-70B

    LanguageOpen weights (restricted use)Open weightsOpen record →

  11. Llama Guard

    LanguageOpen weights (restricted use)Open weightsOpen record →

  12. Llama 2-70B

    LanguageOpen weights (restricted use)Open weightsOpen record →

Show all 52 catalog models

  1. Llama 2-7B

    LanguageOpen weights (restricted use)Open weightsOpen record →

  2. MusicGen

    AudioOpen weights (non-commercial)Open weightsOpen record →

  3. ImageBind

    Multimodal, Vision, Audio, Language, Image generation, SpeechOpen weights (non-commercial)Open weightsOpen record →

  4. Segment Anything Model

    VisionOpen weights (unrestricted)Open weightsOpen record →

  5. LLaMA-65B

    LanguageOpen weights (non-commercial)Open weightsOpen record →

  6. CICERO

    GamesOpen weights (non-commercial)Open weightsOpen record →

  7. Galactica

    Language, BiologyOpen weights (non-commercial)Open weightsOpen record →

  8. EnCodec

    AudioOpen weights (non-commercial)Open weightsOpen record →

  9. Make-A-Video

    VideoHosted access (no API)Closed weightsOpen record →

  10. NLLB

    LanguageOpen weights (unrestricted)Open weightsOpen record →

  11. OPT-175B

    LanguageOpen weights (non-commercial)Open weightsOpen record →

  12. Sparse all-MLP

    LanguageUnreleasedClosed weightsOpen record →

  13. Make-A-Scene

    Image generationUnreleasedClosed weightsOpen record →

  14. data2vec (language)

    LanguageOpen weights (unrestricted)Open weightsOpen record →

  15. data2vec (speech)

    SpeechOpen weights (unrestricted)Open weightsOpen record →

  16. data2vec (vision)

    VisionOpen weights (unrestricted)Open weightsOpen record →

  17. XGLM-7.5B

    LanguageOpen weights (non-commercial)Open weightsOpen record →

  18. Masked Autoencoders ViT-H

    VisionOpen weights (non-commercial)Open weightsOpen record →

  19. XLMR-XXL

    LanguageOpen weights (unrestricted)Open weightsOpen record →

  20. HuBERT

    SpeechOpen weights (unrestricted)Open weightsOpen record →

  21. Generative BST

    LanguageOpen weights (unrestricted)Open weightsOpen record →

  22. wave2vec 2.0 LARGE

    SpeechOpen weights (unrestricted)Open weightsOpen record →

  23. DETR

    VisionOpen weights (unrestricted)Open weightsOpen record →

  24. Transformer - LibriVox + Decoding/Rescoring

    SpeechOpen weights (unrestricted)Open weightsOpen record →

  25. MoCo

    Vision, Image generationOpen weights (non-commercial)Open weightsOpen record →

  26. XLM-RoBERTa

    LanguageOpen weights (non-commercial)Open weightsOpen record →

  27. BART-large

    LanguageOpen weights (unrestricted)Open weightsOpen record →

  28. Pluribus

    GamesUnreleasedClosed weightsOpen record →

  29. FixRes ResNeXt-101 WSL

    VisionOpen weights (non-commercial)Open weightsOpen record →

  30. XLM

    LanguageOpen weights (non-commercial)Open weightsOpen record →

  31. DLRM-2020

    RecommendationUnreleasedClosed weightsOpen record →

  32. ResNeXt-101 Billion-scale

    VisionOpen weights (non-commercial)Open weightsOpen record →

  33. Transformer (Adaptive Input Embeddings) WT103

    LanguageOpen weights (unrestricted)Open weightsOpen record →

  34. ResNeXt-101 32x48d

    VisionOpen weights (non-commercial)Open weightsOpen record →

  35. RetinaNet-R101

    VisionNot documented in sourceWeights unknownOpen record →

  36. RetinaNet-R50

    VisionNot documented in sourceWeights unknownOpen record →

  37. ConvS2S (ensemble of 8 models)

    LanguageNot documented in sourceWeights unknownOpen record →

  38. Mask R-CNN

    VisionNot documented in sourceWeights unknownOpen record →

  39. DTN (Domain Transfer Network)

    Vision, Image generationUnreleasedClosed weightsOpen record →

  40. Character-enriched word2vec

    LanguageNot documented in sourceWeights unknownOpen record →

Filter every reviewed record in the model explorer →

Historical momentum · legacy metric only

+461.0 from first to last sampled observation.

This change is calculated only inside the selected legacy series, from May 8, 2023 to August 29, 2025. The archive crosses methodology changes and cannot be joined to, trended into, or compared numerically with the current snapshot.

First sample827.0llama-13b
Legacy peak1395.92025-05-19 · llama-4-maverick-03-26-experimental
Last sample1288.0llama-4-maverick-17b-128e-instruct
Show all 11 legacy observations
  1. llama-13b827.0
  2. llama-13b865.2
  3. llama-2-13b-chat1029.3
  4. llama-2-13b-chat1027.2
  5. llama-2-13b-chat1024.6
  6. llama-3-70b-instruct1207.2
  7. llama-3.1-405b-instruct1266.2
  8. llama-3.1-405b-instruct-fp81267.2
  9. llama-3.1-405b-instruct-bf161268.0
  10. llama-4-maverick-03-26-experimental1395.9
  11. llama-4-maverick-17b-128e-instruct1288.0

Release record

Latest matched notable models.

Epoch AI records provide dated context only. A release record does not imply a score change.

  1. Muse Spark
  2. Meta's Generative Ads Model (GEM)
  3. Llama 4 Behemoth (preview)
  4. Llama 4 Maverick
  5. Llama 4 Scout

Continue researching

Profiles, guides, and reports.

Evidence boundary

Three sources. Three distinct jobs.

Current frontier: LMArena leaderboard dataset, revision afed939e10281b660a4369206ca505b2bf5e0208, CC BY 4.0.

Historical context: LMArena historical Space artifacts, with one SHA-256 receipt per selected sample and no continuity to the current metric.

Catalog and releases: Epoch AI notable models. No corporate profile is added because our sources do not establish company-level facts beyond model authorship.

Read the full methodology