Jan GGUF models

96 GGUF models in the Jan family, quantized by 30 uploaders. Sizes run from 1.72B to 70.6B parameters. The longest trained context is 256K tokens.

TinySmallMediumLarge
Most downloadedNewestLargestLongest contextSmallest
JA
Jan-v3.5-4B
Jan · janhq
Params4.4B
Context256K
Q4_K_M2.7 GB
GGUF
JA
Jan-v2-VL-high
Jan · janhq
Params8.2B
Context256K
Q4_K_M5.0 GB
Vision
JA
Jan code 4b
Jan · janhq
Params4.4B
Context256K
Q4_K_M2.7 GB
Coding
JA
Jan-Nano-128K
Jan · unsloth
Params4B
Context128K
Q4_K_M2.5 GB
GGUF
JA
Jan v2 VL med
Jan · janhq
Params8.2B
Context256K
Q4_K_M5.0 GB
Vision
JA
Jan-nano-128k
Jan · Menlo
Params4B
Context128K
Q4_K_M2.5 GB
GGUF
JA
Jan nano
Jan · Menlo
Params4B
Context40K
Q4_K_M2.5 GB
GGUF
JA
Jan-v1-4B
Jan · janhq
Params4B
Context256K
Q4_K_M2.5 GB
GGUF
JA
Jan-v3-4b-base-instruct
Jan · janhq
Params4.4B
Context256K
Q4_K_M2.7 GB
Coding
JA
Jan-Nano
Jan · unsloth
Params4B
Context40K
Q4_K_M2.5 GB
GGUF
JA
Jan v3 4B Base Instruct
Jan · bartowski
Params4.41B
Context256K
Q4_K_M2.7 GB
Coding
JA
Jan v1 edge
Jan · janhq
Params1.72B
Context40K
Q4_K_M1.1 GB
GGUF
JA
Jan v1 4B
Jan · bartowski
Params4.02B
Context256K
Q4_K_M2.5 GB
ReasoningTool use
JA
Jan Code 4b
Jan · bartowski
Params4.41B
Context256K
Q4_K_M2.8 GB
CodingTool use
JA
Jan Nano
Jan · bartowski
Params4B
Context40K
Q4_K_M2.5 GB
GGUF
JA
Jan Nano 128k
Jan · bartowski
Params4B
Context128K
Q4_K_M2.5 GB
GGUF
JA
Jan v2 VL High
Jan · mradermacher
Params8.19B
Context256K
Q4_K_M5.0 GB
ReasoningTool useVision
JA
Jan v1 AIO
Jan · prithivMLmods
Params4.02B
Context40K
Q4_K_M1.1 GB
ReasoningTool use
JA
Jan v2 VL Med
Jan · mradermacher
Params8.19B
Context256K
Q4_K_M5.0 GB
ReasoningTool useVision
JA
Jan nano 128k
Jan · gabriellarson
Params4.02B
Context128K
Q4_K_M2.5 GB
ReasoningTool use
JA
Jan v2 VL Med
Jan · mradermacher
Params8.19B
Context256K
Q4_K_M5.0 GB
ReasoningTool useVision
JA
Jan-v2-VL-low
Jan · janhq
Params8.19B
Context256K
Q4_K_M5.0 GB
ReasoningTool useVision
JA
Jan v1 Edge
Jan · bartowski
Params1.72B
Context40K
Q4_K_M1.1 GB
ReasoningTool use
JA
Jan v3 4B Base Instruct
Jan · mradermacher
Params4.41B
Context256K
Q4_K_M2.7 GB
CodingTool use
JA
Jan-v2-VL-max-Instruct-FP8
Jan · fuutott
Params30.5B
Context256K
Q4_K_M18.6 GB
Tool useVision
JA
Jan Nano 128k
Jan · mradermacher
Params4.02B
Context128K
Q4_K_M2.5 GB
ReasoningTool use
JA
Jan Code 4b
Jan · mradermacher
Params4.41B
Context256K
Q4_K_M2.7 GB
CodingTool use
JA
Jan v2 VL High
Jan · mradermacher
Params8.19B
Context256K
Q4_K_M5.0 GB
ReasoningTool useVision
JA
janhq.Jan v1 4B
Jan · DevQuasar
Params4.02B
Context256K
Q4_K_M2.5 GB
ReasoningTool use
JA
Jan v1 4B
Jan · mradermacher
Params4.02B
Context256K
Q4_K_M2.5 GB
ReasoningTool use

About the Jan GGUF models

Jan is represented on ggufmodels.com by 96 GGUF models, quantized by 30 uploaders on Hugging Face. The most downloaded are Jan-v3.5-4B, Jan-v2-VL-high and Jan code 4b. Parameter counts run from 1.72B to 70.6B, and the longest trained context window is 256K tokens.

A Q4_K_M file of a Jan model is between 1.1 to 42.5 GB, which is the quant most people start with. Most files come from mradermacher (31), janhq (9) and RichardErkhov (9). The quants you will meet most often are Q4_K_M (78), Q5_K_M (73), Q6_K (71) and Q3_K_M (69), where the number is how many models offer that quant.

Frequently asked questions

Which Jan GGUF should I download?

The most downloaded Jan GGUF here is Jan-v3.5-4B. Pick the size that fits your GPU first, then choose a quant: Q4_K_M is the usual starting point, and Q5_K_M or Q6_K give more quality if you have memory to spare. The VRAM pages below list what fits each GPU.

How much VRAM does a Jan model need?

Q4_K_M files of Jan models on this site range from 1.1 to 42.5 GB. Add the KV cache for your context length and about 0.5 GB for the runtime, which usually means 1.5 to 3 GB on top of the file at an 8K context.

Where do Jan GGUF files come from?

The most common quantizers here are mradermacher, janhq and RichardErkhov. They convert the original model and publish quantized files on Hugging Face. ggufmodels.com does not host files; every download button opens Hugging Face.

Keep browsing