31 models tracked · known for opt-125m, Llama-3.2-1B-Instruct, Llama-3.1-8B-Instruct
Claims drawn from cited facts, not live model generation.
Meta AI is a lab focused on AI research, open source, and applications. Its blog publishes updates about these same areas of work.
4 cited facts
Meta AI's tracked footprint in the graph includes 31 shipped models and 88 recent blog announcements. That mix signals an actively communicating, research-driven lab — models alongside ongoing public updates, not just product shipping.
2 cited facts
180 benchmark scores across this lab’s models, tagged by evidence status.
| Model | Best score | Claim status |
|---|---|---|
| Llama 3.1-405B | 93.73 | independently reproduced |
| Muse Spark | 88.89 | independently reproduced |
| Llama 2-70B | 87.6 | independently reproduced |
| LLaMA-65B | 86 | self-reported |
| Llama 2-34B | 84.6 | self-reported |
| LLaMA-33B | 83.8 | self-reported |
| Llama 3.1-8B | 82.4 | independently reproduced |
| Llama 3.3 70B | 81.73 | independently reproduced |
| Llama 2-13B | 79.6 | self-reported |
| LLaMA-13B | 77.9 | self-reported |
| Llama 3-8B | 77.07 | independently reproduced |
| Llama 3.2 90B | 73.73 | independently reproduced |
| Llama 2-7B | 73.7 | self-reported |
| Llama 3.1-70B | 73.47 | independently reproduced |
| LLaMA-7B | 73.3 | self-reported |
| Llama 4 Maverick | 73.02 | independently reproduced |
| Llama 3-70B | 72.4 | independently reproduced |
| Llama 4 Scout | 62.27 | independently reproduced |
| Muse Spark 1.1 | 53.32 | unverified |
| opt-125m | — | unverified |
| Llama-3.2-1B-Instruct | — | unverified |
| Llama-3.1-8B-Instruct | — | unverified |
| Meta-Llama-3-8B-Instruct | — | unverified |
| Llama-3.2-1B | — | unverified |
| Meta-Llama-3-8B | — | unverified |
| Llama-3.2-3B-Instruct | — | unverified |
| Llama-2-7b-chat-hf | — | unverified |
| Llama-3.1-70B-Instruct | — | unverified |
Where a model from Meta AI holds the top score, with the benchmark’s integrity grade.
Source facts, citations, and refresh stamp for this record.
Sources: epoch_benchmarks, huggingface_models, lab_blog_rss