Model Spend Arena2ND ED.
12B · MoE · 131,072 ctx · 2026-08-14

JetBrains Mellum 2 12B

Runs on ≤16 GB · Hugging Face ↗

transformersvLLM

Not independently scored by Artificial Analysis

JetBrains' purpose-built coding MoE — 12B total, ~2.5B active, trained on ~10.6T tokens for code completion and infill. The rare coding specialist that actually fits a 16 GB card, and ~2x faster than dense models its size. Custom `MellumForCausalLM` arch.

First-party test · not the AA coding index

BigCodeBench-Hard pass@1 28% (34/121), via local GPU (Ollama GGUF, num_ctx 16384) — comparable full-148, serial. The brutal counterpart to HumanEval — where HumanEval saturates near the top, BCB-Hard spreads the field, so this is the number that actually separates coding ability. A 12B completion/infill specialist — respectable on hard function-completion for its size and blistering speed (~5s/problem end-to-end on a 4060 Ti via Ollama Q4), though it isn't an agentic coder.

Sizes on disk

Real GGUF file sizes = weight VRAM. Add the KV cache for your context (≈2 GB at 32K, fp16).

QuantizationSize
Q3_K_M6.2 GB
IQ4_XS6.8 GB
Q4_07.0 GB
Q4_K_S7.6 GB
Q4_17.7 GB
Q4_K_M8.2 GB
Q48.3 GB
Q5_K_M9.3 GB
Q2_K10.5 GB
Q8_012.9 GB
Q6_K22.1 GB
BF1624.3 GB

Config tips

~8 GB at Q4 — fits an 8-16 GB card. Built for fast completion / fill-in-the-middle rather than long agentic chains, so read a single-shot BCB accordingly. Trending in every "best local coding model" roundup this week.