OPT-175B: AI Model for Research
OPT-175B supports software companies and research labs in performing natural language processing research. It is designed for teams that need a large-scale pretrained model for noncommercial study.
At a glance
- Category
- Software Development
- Best for
- Academic researchers, Industry research labs, Government research organizations, Civil society researchers, Software companies focused on AI research
- Pricing
- Pricing was not clearly available from the provided evidence. The model is released under a noncommercial license for research purposes. Buyers should confirm current pricing on the vendor website.
- Key use cases
- NLP Research, AI Safety Study, Model Scaling Analysis, Reproducible AI Research
OPT-175B is a large-scale language model released by Meta AI. It consists of 175 billion parameters and is designed for the scientific and research community to study how large language models work and to explore methods for mitigating issues like bias and toxicity.
The model is intended for academic researchers, government agencies, civil society, and industry research laboratories. It provides both the pretrained models and the training code, which differs from many large models available only via paid APIs.
It supports research into natural language processing (NLP) and AI behavior. The release also includes a suite of smaller baseline models to help researchers study the impact of model scale on performance.
Buyers and researchers should confirm that their intended use falls under the noncommercial license agreement before requesting access.
Key Features
A high-capacity language model trained on publicly available data sets.
Includes the codebase needed to train and deploy the model.
Provides a suite of smaller models ranging from 125M to 30B parameters for scale-based research.
Access is granted for research use cases to focus on research applications.
Includes a logbook documenting the training process and compute usage.
Use Cases
Studying natural language processing capabilities and behaviors using a large-scale model.
Analyzing potential harms and designing detection and mitigation strategies for bias and toxicity.
Comparing the 175B model with smaller baseline models to understand the effect of parameter scale.
Using shared pretrained models and code to conduct research that other labs can replicate.
Best For
Pricing
Pricing was not clearly available from the provided evidence. The model is released under a noncommercial license for research purposes. Buyers should confirm current pricing on the vendor website.
FAQ
Access is granted to academic researchers, government organizations, civil society, and industry research laboratories.
No, the model is released under a noncommercial license specifically to focus on research use cases.
The release includes the 175-billion-parameter pretrained model, the code needed to train and use it, and several smaller baseline models.
Source category: Software Development
Source subcategory: Machine Learning Platform
