Mistral's Ministral 3 14B dense model with vision encoder. The largest model in the Ministral 3 family, Ministral 3 14B offers frontier capabilities and performance comparable to its larger Mistral Small 3.2 24B counterpart. A powerful and efficient language model with vision capabilities. This model is the instruct post-trained version, fine-tuned for instruction tasks, making it ideal for chat and instruction based use cases.
On-demand DeploymentDocs | On-demand deployments allow you to use Ministral 3 14B Instruct 2512 on dedicated GPUs with Fireworks' high-performance serving stack with high reliability and no rate limits. |
