Run ChindaLLM 4B with Ollama

By the end, ChindaLLM 4B answers in Thai on your own machine, in the terminal and through a local API.

You need macOS, Linux or Windows, room for a 2.5 GB download, and at least 4 GB of RAM (8 GB recommended).

1. Install Ollama​

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

On Windows, use the installer from ollama.com/download; on any system, ollama --version confirms the install.

2. Download the model​

ollama pull iapp/chinda-qwen3-4b

The download is about 2.5 GB; when it finishes, ollama list shows iapp/chinda-qwen3-4b:latest.

3. Chat in the terminal​

# a chat session: type at the >>> prompt, /bye to leave
ollama run iapp/chinda-qwen3-4b

# one question, then exit
ollama run iapp/chinda-qwen3-4b "ช่วยแก้สมการ 2x + 5 = 15 ให้หน่อย"

4. Call it from your code​

The local API listens on http://localhost:11434; if it is not running, start it with ollama serve.

curl http://localhost:11434/api/chat -d '{
  "model": "iapp/chinda-qwen3-4b",
  "messages": [
    {"role": "user", "content": "อธิบายเกี่ยวกับปัญญาประดิษฐ์ให้ฟังหน่อย"}
  ]
}'

import requests

def chat_with_chinda(message):
    url = "http://localhost:11434/api/generate"
    data = {
        "model": "iapp/chinda-qwen3-4b",
        "prompt": message,
        "stream": False
    }
    response = requests.post(url, json=data)
    return response.json()["response"]

print(chat_with_chinda("สวัสดีครับ"))

Chinda 4B answers best with the relevant text in the prompt; asked for facts on their own, such as recent news or statistics, it can give wrong answers.

Benchmarks and licence: ChindaLLM 4B · On Ollama: iapp/chinda-qwen3-4b · Weights: Hugging Face