Mistral AI
Apache 2.0 licensed 24-billion-parameter model built for low latency and self-hosted deployment.
Mistral Small 3 is a 24-billion-parameter open-weight model released under Apache 2.0 and optimised for latency. Mistral positions it against considerably larger models, and the 3.1 release adds vision understanding and a 128,000-token context window.
Workstation or edge server with sufficient memory
Specifications taken from current manufacturer documentation and last checked on 2026-09-08.
Source: manufacturer documentation
Specifications are summarised from publicly published manufacturer material and are provided for orientation only. Always confirm current figures directly with the manufacturer before purchasing or designing in.
MIT-licensed runtime for pulling and serving open-weight models locally through a CLI and local API.
MIT-licensed C and C++ inference engine for open-weight language models, on CPU or GPU backends.
MIT-licensed 14-billion-parameter small language model built for low-latency reasoning tasks.