mannix/jan-nano:ud_q8_k

mannix/ jan-nano:ud_q8_k_xl

1,683 Downloads Updated 7 months ago

Jan-Nano is a compact 4-billion parameter language model specifically designed and trained for deep research tasks.

tools thinking

ollama run mannix/jan-nano:ud_q8_k_xl

curl http://localhost:11434/api/chat \
  -d '{
    "model": "mannix/jan-nano:ud_q8_k_xl",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from ollama import chat

response = chat(
    model='mannix/jan-nano:ud_q8_k_xl',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

const response = await ollama.chat({
  model: 'mannix/jan-nano:ud_q8_k_xl',
  messages: [{role: 'user', content: 'Hello!'}],
})
console.log(response.message.content)

Details

Updated 7 months ago

7 months ago

5c795ee6fbfb · 5.1GB ·

model

archqwen3

parameters4.02B

quantizationQ8_0

5.1GB

template

{{- $lastUserIdx := -1 -}} {{- range $idx, $msg := .Messages -}} {{- if eq $msg.Role "user" }}{{ $la

1.7kB

params

{ "min_p": 0, "repeat_penalty": 1, "stop": [ "<|im_start|>", "<|im_end|>

129B

Readme

https://huggingface.co/Menlo/Jan-nano

You can hide the thinking, without disabling it, using the CLI with the switch --hidethinking

Optionally, use the /no_think in the prompt or system prompt to disable thinking.