Neura Market
StableLM-3B logo

StableLM-3B

Free

A 3B parameter causal language model for text generation, open and customizable.

FreeFree tier
Inputs: textOutputs: text
Type
Open Source
Company
Stability AI
Visit Website GitHub (38 stars)

About StableLM-3B

StableLM-3B is a 3-billion parameter causal language model developed by Stability AI, designed for text generation tasks. It is trained on a diverse mixture of datasets including Falcon RefinedWeb, RedPajama, Pile v2, StarCoder, and peS2o. The model is released under the CC-BY-SA-4.0 license and supports the Transformers library with SafeTensors format. It achieves competitive performance on benchmarks such as HellaSwag (75.94% accuracy) and Winogrande (71.19% accuracy), though it shows limitations in mathematical reasoning (GSM8k: 3.34%) and truthfulness (TruthfulQA: 37.2%). The model is available for download on Hugging Face and can be fine-tuned for custom applications.

Key Features

3 billion parameters for efficient text generation
Trained on diverse datasets including Falcon RefinedWeb, RedPajama, Pile v2, StarCoder, and peS2o
Released under CC-BY-SA-4.0 license
Supports Hugging Face Transformers and SafeTensors format
Causal language model (causal-lm) architecture
Open source and free to use, modify, and distribute

Pros & Cons

Pros
  • Open source with permissive CC-BY-SA-4.0 license
  • Efficient 3B parameter size balances performance and resource requirements
  • Trained on a broad mix of high-quality datasets
  • Strong performance on commonsense reasoning (HellaSwag) and coreference resolution (Winogrande)
  • Easily integrated with the Transformers library and inference endpoints
Cons
  • Poor mathematical reasoning (GSM8k accuracy only 3.34%)
  • Low truthfulness score (TruthfulQA 37.2%) indicating potential for generating inaccurate facts
  • Smaller than many state-of-the-art models, limiting performance on complex tasks
  • Requires fine-tuning for specialized applications beyond basic text generation

Best For

Text generation and completionFine-tuning for domain-specific language tasksConversational AI and chatbot developmentContent creation and creative writing assistanceResearch and experimentation in natural language processing

FAQ

What is StableLM-3B?
StableLM-3B is a 3-billion parameter causal language model developed by Stability AI, optimized for text generation tasks.
What license does StableLM-3B use?
The model is released under the Creative Commons Attribution-ShareAlike 4.0 International (CC-BY-SA-4.0) license.
What are the benchmark scores?
On the Open LLM Leaderboard, the model achieves: ARC (25-shot) 46.59%, HellaSwag (10-shot) 75.94%, MMLU (5-shot) 45.23%, TruthfulQA (0-shot) 37.2%, Winogrande (5-shot) 71.19%, and GSM8k (5-shot) 3.34%.
Can I use StableLM-3B commercially?
Yes, the CC-BY-SA-4.0 license allows commercial use, but any derivative works must be shared under the same license.