AI21/Mixture of Experts

AJamba 2 Mini

Name: Jamba 2 Mini
Author: AI21

Jamba2 Mini is an open source small language model built for enterprise reliability. With 12B active parameters (52B total), it delivers precise question answering without the computational overhead of reasoning models.

chatreasoningtool_usemultilingual

52B

Parameters (12B active)

256K

Context length

Benchmarks

Quantizations

Architecture

MoE

Released

2026-02-15

Layers

KV Heads

Head Dim

128

Family

jamba

Quantization Options

Quant	Bits	VRAM	Quality
IQ3_XXS	3.25	21.6 GB	low
IQ3_XS	3.5	23.2 GB	low
Q3_K_S	3.64	24.1 GB	low
IQ3_M	3.76	24.9 GB	low
Q3_K_M	4	26.5 GB	low
Q3_K_L	4.3	28.4 GB	moderate
IQ4_XS	4.46	29.5 GB	moderate
Q4_K_S	4.67	30.8 GB	moderate
Q4_K_M	4.89	32.3 GB	good
Q5_K_S	5.57	36.7 GB	good
Q5_K_M	5.7	37.5 GB	good
Q6_K	6.56	43.1 GB	excellent
Q8_0	8.5	55.7 GB	lossless
FP16	16	104.5 GB	lossless

Select your GPU above to see speed estimates and compatibility for each quantization.

▸ READY TO RUN THIS?RENT BY THE HOUR

RENT A GPU AND RUN JAMBA 2 MINI NOW

Rent on RunPod →Or Vast.ai →

Spin up an A100 / H100 / 4090 in ~60s. Pay by the second. Cancel anytime.

Community Ratings

Loading ratings...

Benchmarks (6)

IFEval20.3

MMLU-PRO16.6

BBH10.7

MUSR3.7

GPQA2.5

MATH1.6

Run this model

▸Easiest way to get started·Beginners

DOCS ↗

curl -fsSL https://ollama.com/install.sh | sh

$ollama run jamba:52b-q4_K_M

Tag may need adjustment — check ollama.com/library/jamba for available tags.

▸ SETUP GUIDE

Auto-setup with fitmyllm CLI

Detects your GPU, recommends the best model, downloads it, and starts chatting — zero config. Benchmarks your speed and contributes anonymous data to improve predictions.

pip install fitmyllmthen run fitmyllmLearn more

Auto-detect GPULive tok/s in chatSpeed benchmarks9 inference engines

HuggingFace GGUF Downloads Build Hardware