MIST
Our flagship large language model family. Built with advanced merging (DARE-TIES, frankenmerging) and GRPO reasoning tuning on the Llama 3.1 lineage, spanning 8B, 70B, and 140B parameters, with 4-bit quantised releases for accessible deployment.
MISTMIST-Mini-8B
General-purpose 8B merged LLM built on the Llama 3.1 lineage with advanced model merging.
MISTMIST-1-70B
General-purpose 70B merged LLM built on the Llama 3.1 lineage with advanced model merging.
MISTMIST-1-140B
General-purpose 140B frankenmerge LLM built on the Llama 3.1 lineage with advanced model merging.
MISTMIST-1-140B-4bit
4-bit NF4 quantised release of MIST-1-140B: frontier-scale weights on accessible hardware.
MISTMIST-Mini-8B-Thinking
Reasoning-tuned MIST model trained with GRPO to think step-by-step before answering.
Explore Other Families
More of the stack, organised the same way.
LID
Language identification: from 5 Nigerian languages to 25 languages worldwide.
DiacNet
Diacritic & tone restoration for YorĂčbĂĄ, Igbo, Hausa and beyond.
Prism
Vision models: super-resolution upscaling, denoising, and image steganography.