JetBrains releases Mellum2.1, a 12B MoE open model for coding agents

JetBrains has released Mellum2.1, an open 12B mixture-of-experts model with 2.5B active parameters under Apache 2.0. Reinforcement learning in real repositories lifted SWE-bench Verified from 2.0 to 47.0, while LiveCodeBench v6 reached 82.0.

• 12B total and 2.5B active parameters, 64 experts, 131,072-token context
• SWE-bench Verified rose from 2.0 to 47.0 after RL in real repos
• LiveCodeBench v6 score of 82.0 vs 75.4 for Qwen3.5-9B
• GGUF builds start at 7.0 GB for llama.cpp, Ollama and LM Studio