MIT Licensed 400 Billion Parameter model by Jabir Project

tools

ollama run haghiri/jabir-400b:q4_k

curl http://localhost:11434/api/chat \
  -d '{
    "model": "haghiri/jabir-400b:q4_k",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from ollama import chat

response = chat(
    model='haghiri/jabir-400b:q4_k',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

const response = await ollama.chat({
  model: 'haghiri/jabir-400b:q4_k',
  messages: [{role: 'user', content: 'Hello!'}],
})
console.log(response.message.content)

Applications

Claude Code ollama launch claude --model haghiri/jabir-400b:q4_k

Codex ollama launch codex --model haghiri/jabir-400b:q4_k

OpenCode ollama launch opencode --model haghiri/jabir-400b:q4_k

OpenClaw ollama launch openclaw --model haghiri/jabir-400b:q4_k

Models

View all →

Name

1 model

Size

Context

Input

jabir-400b:q4_k

149GB · 128K context window · Text · 1 year ago

jabir-400b:q4_k

149GB

128K

Text

Readme

Jabir 400B

Introduction

This model is part of Jabir Project by Muhammadreza Haghiri as a large and well trained model on multiple languages with a goal of democratizing AI.

After the success of other Haghiri’s project, Mann-E he gathered a team to work on a large language model which has a good understanding of Persian language (alongside other languages).

Jabir is available in 405 billion parameters, based on LLaMA 3.1 and it wasn’t openly available until now. The project will continue to improve and this repository will be updated as well.

Benchmarks

License

This model has been released under MIT License.