Jurassic-1 Models by AI21 Labs
About
Jurassic-1, a family of large language models developed by AI21 Labs, includes two notable models: J1-Jumbo, boasting 178 billion parameters, and J1-Large, with 7 billion parameters. These autoregressive models excel in predicting the next word in a sequence, enabling them to tackle various tasks such as text generation, question answering, and text classification. Trained on an extensive corpus of English text, Jurassic-1 models are renowned for their impressive vocabulary of 250,000 tokens, which enhances efficiency and reduces latency compared to other LLMs. Access to these models is provided through AI21 Studio, an API, and interactive web platform that also allows for custom model training with minimal datasets.
Current Variants
Use-when guidance is derived from seed capabilities, context, release, and replacement fields.
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| J1-Jumbo | Use when the workload needs 178B parameters. | 2021-08 | 178B parameters | Current |
| J1-Grande | Use when the workload needs 17B parameters. | 2021-08 | 17B parameters | Current |
| J1-Large | Use when the workload needs 7.5B parameters. | 2021-08 | 7.5B parameters | Current |
Release Timeline
1 release groupSpecifications(3 models)
Frequently Asked Questions
- What is Jurassic-1 used for?
- Jurassic-1 is used for chatbot and role-playing use cases. The family description and listed model capabilities point to those workloads as the best fit.
- How does Jurassic-1 compare to Jamba 1.7?
- Jurassic-1 by AI21 Labs is strongest where you need chatbot and role-playing use cases, while Jamba 1.7 by AI21 Labs is the closest related family to check for agent workflows and tool use. Jurassic-1 has 3 listed variants, while Jamba 1.7 reaches up to 256k context, so compare the specs and pricing tables before choosing a production model.
- Which Jurassic-1 model should I use?
- If price is the main constraint, use the pricing table first because Jurassic-1 does not have complete provider pricing in the local data. For the most capable/latest local choice, evaluate J1-Jumbo.
