bjoernb/gemma4-26b-think

bjoernb/ gemma4-26b-think

2,750 Downloads Updated 4 months ago

Gemma 4 26B MoE (Google DeepMind) with thinking mode enabled. Mixture-of-Experts — 25.2B total / 3.8B active parameters, 256K context. Supports text and image input. Knowledge cutoff: January 2025.

vision tools thinking

ollama run bjoernb/gemma4-26b-think

curl http://localhost:11434/api/chat \
  -d '{
    "model": "bjoernb/gemma4-26b-think",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from ollama import chat

response = chat(
    model='bjoernb/gemma4-26b-think',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

const response = await ollama.chat({
  model: 'bjoernb/gemma4-26b-think',
  messages: [{role: 'user', content: 'Hello!'}],
})
console.log(response.message.content)