ornith

715.3K Downloads Updated 4 months ago

A self-improving family of open-source models for agentic coding

tools thinking 9b 35b

ollama run ornith

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

from ollama import chat

response = chat(
    model='ornith',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

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

Applications

Claude Code ollama launch claude --model ornith

OpenCode ollama launch opencode --model ornith

Hermes Agent ollama launch hermes --model ornith

OpenClaw ollama launch openclaw --model ornith

Models

View all →

Name

9 models

Size / Usage

Context

Input

ornith:latest

5.6GB · 256K context window · Text · 4 months ago

ornith:latest

5.6GB

256K

Text

ornith:9b

latest

5.6GB · 256K context window · Text · 4 months ago

ornith:9b latest

5.6GB

256K

Text

ornith:35b

21GB · 256K context window · Text · 4 months ago

ornith:35b

21GB

256K

Text

Readme

Highlights:

State-of-the-Art Coding Agents: Achieving state-of-the-art performance among open-source models of comparable size on coding benchmarks such as Terminal-Bench 2.1, SWE-Bench, NL2Repo and OpenClaw.
Self-Improving Training Framework: Ornith-1.0 employs RL to learn to generate not only solution rollouts, but also the scallfold that drive those rollouts. By jointly optimizing the scaffold and the resulting solution, the model discovers better search trajectories and generates higher-quality solutions.
Licence: MIT licensed, globally accessible, and free from regional limitations.

9B

ollama run ornith:9b

35B

ollama run ornith:35b

References

Blog