AI21 Labs/Mixture of Experts

AJamba 1.5 Mini 52B

Name: Jamba 1.5 Mini 52B
Author: AI21 Labs

AI21's Jamba 1.5 Mini — hybrid SSM-Transformer MoE. Fast with 256K context window.

chatreasoningThinkingTool Use

51.6B

Parameters (12B active)

256K

Context length

Benchmarks

Quantizations

20K

HF downloads

Architecture

MoE

Released

2024-08-22

Layers

KV Heads

Head Dim

128

Family

jamba

Quantization Options

Quant	Bits	VRAM	Quality
IQ3_XXS	3.25	21.5 GB	low
IQ3_XS	3.5	23.1 GB	low
Q3_K_S	3.64	24.0 GB	low
IQ3_M	3.76	24.7 GB	low
Q3_K_M	4	26.3 GB	low
Q3_K_L	4.3	28.2 GB	moderate
IQ4_XS	4.46	29.3 GB	moderate
Q4_K_S	4.67	30.6 GB	moderate
Q4_K_M	4.89	32.0 GB	good
Q5_K_S	5.57	36.4 GB	good
Q5_K_M	5.7	37.3 GB	good
Q6_K	6.56	42.8 GB	excellent
Q8_0	8.5	55.3 GB	lossless
FP16	16	103.7 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 1.5 MINI 52B 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 (14)

HumanEval64.0

IFEval64.0

MATH57.0

MMLU-PRO42.0

MATH-50035.7

GPQA Diamond30.2

BBH10.7

AA Intelligence8.0

SciCode8.0

LiveCodeBench6.2

HLE5.1

MUSR3.7

GPQA2.5

AIME1.0

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