inkling-2-small

60 Downloads Updated 11 hours ago

594B Mixture-of-Experts model (17B active) with image input and up to 1M tokens of context, built for agentic coding and knowledge work.

vision tools thinking cloud

Cost /1M tokens

$0.30

input

$0.03

cached

$1.00

output

Context

1M tokens

Size

594B parameters

ollama run inkling-2-small:cloud

curl https://ollama.com/api/chat \
  -H "Authorization: Bearer $OLLAMA_API_KEY" \
  -d '{
    "model": "inkling-2-small",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

import os
from ollama import Client

client = Client(
    host='https://ollama.com',
    headers={'Authorization': 'Bearer ' + os.environ['OLLAMA_API_KEY']},
)

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

import { Ollama } from 'ollama'

const ollama = new Ollama({
  host: 'https://ollama.com',
  headers: { Authorization: 'Bearer ' + process.env.OLLAMA_API_KEY },
})

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

Applications

Claude Code ollama launch claude --model inkling-2-small:cloud

OpenCode ollama launch opencode --model inkling-2-small:cloud

Hermes Agent ollama launch hermes --model inkling-2-small:cloud

OpenClaw ollama launch openclaw --model inkling-2-small:cloud

Models

View all →

Name

1 model

Size / Usage

Context

Input

inkling-2-small:cloud

- · 1M context window · Text, Image · 11 hours ago

inkling-2-small:cloud

Text, Image

Readme

Inkling 2 Small is a Mixture-of-Experts model from Thinking Machines Lab, with 594B total and 17B active parameters. It reads text and images, supports variable reasoning effort, and is built for agentic coding, research and knowledge work.

Get started

ollama run inkling-2-small:cloud

Key Features

Efficient Mixture-of-Experts: 594B parameters, with each token routed to 8 of 384 experts (17B active).
Agentic coding and research: built for long-horizon coding and autoresearch tasks.
Vision: reads documents, charts, screenshots and images, including OCR and visual reasoning.
Knowledge work: looks up information, calls tools over MCP and produces finished deliverables.
Long context: up to 1M tokens.

Benchmarks

Thinking Machines reports 86.3% on SWE-bench Pro v2 (Hard), 72.1% on DeepSWE, 85.4 on MCP Atlas and 81.6% on MMMU Pro. See the model card for details.

Reference

Thinking Machines launch post · Model card · Hugging Face card

### Get started
```shell
ollama run inkling-2-small:cloud
```

### Key Features
- **Efficient Mixture-of-Experts**: 594B parameters, with each token routed to 8 of 384 experts (17B active).
- **Agentic coding and research**: built for long-horizon coding and autoresearch tasks.
- **Vision**: reads documents, charts, screenshots and images, including OCR and visual reasoning.
- **Knowledge work**: looks up information, calls tools over MCP and produces finished deliverables.
- **Long context**: up to 1M tokens.

### Benchmarks
Thinking Machines reports 86.3% on SWE-bench Pro v2 (Hard), 72.1% on DeepSWE, 85.4 on MCP Atlas and 81.6% on MMMU Pro. See the [model card](https://thinkingmachines.ai/model-card/inkling-2-small/) for details.

### Reference
[Thinking Machines launch post](https://thinkingmachines.ai/news/inkling-2-small/) · [Model card](https://thinkingmachines.ai/model-card/inkling-2-small/) · [Hugging Face card](https://huggingface.co/thinkingmachines/Inkling-2-Small)

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