JetBrains/Mellum2-12B-A2.5B-Thinking — 12B / 2.5B active · MOE · 128K ctx
JetBrains’ reasoning-augmented code MoE (12B total / 2.5B active) that emits explicit <think> chains for debugging, planning, and agentic coding
AI Model Collection
Summary
Mellum 2 is a JetBrains collection of open-source 12B parameter text generation models on Hugging Face. It includes six variants like Thinking and Instruct, updated recently with visible engagement stats.
Highlights
- Six variants including Thinking, Instruct, and Base models.
- All models use the 12B parameter Mellum2-12B-A2.5B architecture.
- Hosted on Hugging Face by JetBrains for open access.
- Includes metadata like download counts and update times.
- Focused on text generation tasks.
auto-generated
Context
Audience
Machine Learning Engineers
DomainMachine Learning
Formatmodel repository
Accessopen source
Source Overview
Mellum 2 is a JetBrains collection of Mellum2 12B A2.5B text generation model weights hosted on Hugging Face.
It includes variants for Thinking, Instruct, Thinking-SFT, Instruct-SFT, Base, and Base-Pretrain, all with 12B parameters. Each model card lists recent update times and basic usage statistics such as downloads and likes.
Topics
Related
Hugging FaceJetBrains AILarge Language ModelsText GenerationModel Weights
