Pick up to four organizations and see side by side what each one publishes. The page address keeps your selection, so the link can be shared.
| Measure | |||
|---|---|---|---|
| Country | 🇺🇸 United States | 🇺🇸 United States | 🇨🇳 China |
| Type | Corporate lab | Corporate lab | Corporate lab |
| Location | San Francisco, California | San Francisco, California | Hangzhou |
| Models on Hugging Face | – | 39 | 106 |
| Datasets on Hugging Face | – | 16 | 2 |
| Downloads in 30 days | – | 58.8M | 24M |
| Followers | 5.9K | 45K | 155K |
| Likes | – | 38K | 89K |
| Releases in 90 days | – | 0 | 3 |
| Last upload | – | 2 months ago | 10 days ago |
| Smallest model | – | 38M | 673M |
| Largest model | – | 116.8B | 1.65T |
| Models offered through an API | 16 | 62 | 14 |
| Latest API model | Claude Fable 5.2 (11 days ago) | GPT-6.2 Astra Pro (3 days ago) | DeepSeek V4.1 Flash 1017 (1 month ago) |
| Largest API context | 1M tokens | 1.1M tokens | 1M tokens |
| Input price per million tokens | US$ 0.1 – US$ 15 | US$ 0.02 – US$ 30 | US$ 0.04 – US$ 1.32 |
| Papers on arXiv | – | – | – |
| AI works since 2024 (OpenAlex) | 347 | 158 | – |
| Most used licenses | – | apache-2.0, mit, cc-by-sa-4.0 | other, mit, apache-2.0 |
| Activity over 8 quarters |
The highest value in each row is highlighted. The numbers count public volume, not quality. API models, context and prices come from OpenRouter's public catalog: they show what is offered there, not each lab's full catalog. arXiv papers are only attributed for organizations in the Brazilian research; for the others the row is empty.