NVIDIA has released “Mistral-NeMo-Minitron 8 billion,” a compact language model of the Mistral NeMo 12 billion model that is accurate and compute-efficient.

The model includes AI-powered chatbots, virtual assistants, content generators, and educational tools. It is also capable of language understanding, common sense reasoning, mathematical reasoning, summarization, coding, and generating truthful answers.

The model is available as an NVIDIA NIM microservice with a standard API.

The company claims the model has comparable accuracy to the original model at lower computational cost with compute efficiency across GPU-accelerated data centers, clouds, and workstations.

Analyst QuickTake: NVIDIA has been releasing models frequently over the past few months. In July 2024, it partnered with Mistral AI to release Mistral NeMo 12B, a 12-billion-parameter multilingual language model. This week, it released StormCast, a GenAI model for high-fidelity atmospheric dynamics, and Llama-3.1-Minitron 4 billion, a compressed version of the Llama 3 model, to run on resource-constrained devices.