Upstage, a South Korean AI-based startup, has released the preview of its large language model (LLM), Solar Pro.
The model is a 22-billion-parameter model optimized to run on a single GPU. Additionally, it features a depth-up scaling method, an advanced data recipe, English-language input support, and a context window of 4,096 tokens.
Furthermore, the preview is an open-source model with free API access until its official release in November.
The company claims the model outperforms Microsoft’s Phi 3 Medium, Meta's Llama 3.1 8B, Mistral NeMo 12B, and Google's Gemma 2 27B. Moreover, the company also claims it can automate and streamline tasks across industries such as healthcare, finance, and legal services, enhancing productivity and efficiency.
By using this site, you agree to allow SPEEDA Edge and our partners to use cookies for analytics and personalization. Visit our privacy policy for more information about our data collection practices.