ROKTEAM/ rok-oss:latest

223 Downloads Updated 3 months ago

ROK‑OSS is a fast, general-purpose AI assistant built on LLaMA 4. It provides concise, structured outputs for coding, writing, analysis, and real-world workflows. Designed for clarity and minimal fluff.

tools 3b

ollama run ROKTEAM/rok-oss

curl http://localhost:11434/api/chat \
  -d '{
    "model": "ROKTEAM/rok-oss",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from ollama import chat

response = chat(
    model='ROKTEAM/rok-oss',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

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

Details

Updated 3 months ago

3 months ago

09b88b928f21 · 4.9GB ·

model

archllama

parameters8.03B

quantizationQ4_K_M

4.9GB

license

LLAMA 3.1 COMMUNITY LICENSE AGREEMENT Llama 3.1 Version Release Date: July 23, 2024 “Agreement”

12kB

system

You are ROK-OSS, a fast, general-purpose AI assistant. You provide clear, concise answers, structure

213B

params

{ "num_ctx": 65536, "repeat_penalty": 1.05, "stop": [ "<|start_header_id|>",

165B

template

<|system|> {{ .System }} <|user|> {{ .Prompt }} <|assistant|>

73B

Readme

ROK‑OSS is a fast, general-purpose AI assistant designed for developers, students, and anyone needing a versatile helper. It generates code, produces structured outputs, summarizes text, answers questions, and assists with workflow automation. Outputs are concise, predictable, and tool-friendly.

Performance Comparison

Bar graph illustrates approximate performance on coding, reasoning, and summarization tasks compared to GPT‑4 and LLaMA 2.

Sources:

GPT‑4 vs LLaMA 2 benchmarks: codesmith.io

LLaMA 4 community evaluations: abhs.in

LLaMA 2 summarization: anyscale.com

Installation / Run Locally

Pull from Ollama

ollama pull ROKTEAM/rok-oss

Run

ollama run ROKTEAM/rok-oss

Notes

Outputs follow the ROK‑OSS style: concise, structured, predictable.
Optimized for general-purpose assistant tasks, not specialized reasoning benchmarks.
Bar graph and benchmarks are illustrative based on available community evaluations.

![New Project (19).png](/assets/ROKTEAM/rok-oss/afcafc72-8917-4d63-a766-0000431123a5)

Performance Comparison

Bar graph illustrates approximate performance on coding, reasoning, and summarization tasks compared to GPT‑4 and LLaMA 2.

![Untitled design (1).png](/assets/ROKTEAM/rok-oss/04742eaa-4977-4a6e-8962-23d8e20efd22)

Sources:

GPT‑4 vs LLaMA 2 benchmarks: [codesmith.io](https://www.codesmith.io/blog/meta-llama-2-vs-gpt-4-which-ai-model-comes-out-on-top?utm_source=chatgpt.com)

LLaMA 4 community evaluations: [abhs.in
](https://www.abhs.in/blog/meta-llama-4-multimodal-open-source-benchmarks-2026?utm_source=chatgpt.com)

LLaMA 2 summarization: [anyscale.com](https://www.anyscale.com/blog/llama-2-is-about-as-factually-accurate-as-gpt-4-for-summaries-and-is-30x-cheaper?utm_source=chatgpt.com)

# Installation / Run Locally
# Pull from Ollama
`ollama pull ROKTEAM/rok-oss`

# Run
`ollama run ROKTEAM/rok-oss`

Notes

- Outputs follow the ROK‑OSS style: concise, structured, predictable.

- Optimized for general-purpose assistant tasks, not specialized reasoning benchmarks.

- Bar graph and benchmarks are illustrative based on available community evaluations.

Paste, drop or click to upload images (.png, .jpeg, .jpg, .svg, .gif)

ROK‑OSS is a fast, general-purpose AI assistant built on LLaMA 4. It provides concise, structured outputs for coding, writing, analysis, and real-world workflows. Designed for clarity and minimal fluff.

Details

Readme

Installation / Run Locally

Pull from Ollama

Run

ROK‑OSS is a fast, general-purpose AI assistant built on LLaMA 4. It provides concise, structured outputs for coding, writing, analysis, and real-world workflows. Designed for clarity and minimal fluff.