bonsai-2

155 Downloads Updated 3 hours ago

Bonsai 2 from Prism ML is a family of vision models built for ternary weights – smaller models with close to full-precision quality.

vision tools thinking 2b 4b 9b 27b

ollama run bonsai-2

curl http://localhost:11434/api/chat \
  -d '{
    "model": "bonsai-2",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from ollama import chat

response = chat(
    model='bonsai-2',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

const response = await ollama.chat({
  model: 'bonsai-2',
  messages: [{role: 'user', content: 'Hello!'}],
})
console.log(response.message.content)

Applications

Claude Code ollama launch claude --model bonsai-2

OpenCode ollama launch opencode --model bonsai-2

Hermes Agent ollama launch hermes --model bonsai-2

OpenClaw ollama launch openclaw --model bonsai-2

Models

View all →

Name

9 models

Size / Usage

Context

Input

bonsai-2:latest

3.2GB · 256K context window · Text, Image · 3 hours ago

bonsai-2:latest

3.2GB

256K

Text, Image

bonsai-2:2b

911MB · 256K context window · Text, Image · 3 hours ago

bonsai-2:2b

911MB

256K

Text, Image

bonsai-2:4b

1.6GB · 256K context window · Text, Image · 3 hours ago

bonsai-2:4b

1.6GB

256K

Text, Image

bonsai-2:9b

latest

3.2GB · 256K context window · Text, Image · 3 hours ago

bonsai-2:9b latest

3.2GB

256K

Text, Image

bonsai-2:27b

8.3GB · 256K context window · Text, Image · 3 hours ago

bonsai-2:27b

8.3GB

256K

Text, Image

Readme

Bonsai 2 from Prism ML is a family of vision models built for ternary weights (−1, 0 or +1): about half the size of standard 4-bit models, at close to full-precision quality.

All sizes support thinking, tool calling, image input and a 262K-token context window.

Bonsai 2 from Prism ML is a family of vision models built for ternary weights – smaller models with close to full-precision quality.

Applications

Models

Readme

References