Particle 1
Neural Network
Particle 1 is a Particle AI agentic model for coding: codebase edits and terminal work with efficient token usage.
Specifications
- Input
- text, images, documents
- Output
- text
- Context size
- 1 048 576 tokens
- Max answer length
- 131 072 tokens
- Reasoning effort
- none, minimal, low, medium, high, extra high, max
- Capabilities
- reasoning, function calling
Prompt cost
(per 1M tokens)
Answer cost
(per 1M tokens)
import OpenAI from 'openai';
const openai = new OpenAI({
apiKey: '<your bothub access token>',
baseURL: 'https://openai.bothub.chat/v1'
});
// Sync - Text generation
async function main() {
const chatCompletion = await openai.chat.completions.create({
messages: [{ role: 'user', content: 'Say this is a test' }],
model: 'particle-1',
});
}
// Async - Text generation
async function main() {
const stream = await openai.chat.completions.create({
messages: [{ role: 'user', content: 'Say this is a test' }],
model: 'particle-1',
stream: true
});
for await (const chunk of stream) {
const part: string | null = chunk.choices[0].delta?.content ?? null;
}
}
main();
How it works Particle 1?
Particle 1 is an agentic model for programming from Particle AI, fine-tuned for engineering tasks. It is designed for terminal work, targeted changes in large codebases, and data processing, while efficient token usage reduces the cost of long sessions. Through BotHub, you can connect it without a VPN or foreign card, pay in rubles as you go, and manage other models in one window, switching between them without rewriting integrations. For building multi-model systems, there is a unified OpenAI-compatible API. Requests are encrypted via AES-GCM and not stored. Teams have access to contracts, invoices, electronic document management (EDM), an admin panel, and limits. Particle 1 is useful for refactoring across multiple files, analyzing third-party repositories to find the cause of test failures, automating console routines, writing tests, preparing data migrations, or building an agent that calls tools and verifies results.Frequently asked questions about Particle 1
You can use generated results for commercial purposes. You own all rights to the content you create. The only restriction: make sure your prompt does not include copyrighted third-party material. You are responsible for respecting the rights to any input data.
particle-1 is a text model for working with large volumes of information: documents, images, code. It reasons before answering, calls functions, and returns structured JSON, making it suitable for documentation analysis, code review, debugging, multi-step calculations, and generating answers based on your materials.
In a single response, particle-1 outputs up to 131,072 tokens — a large volume of text, code, or data tables. The model's context window reaches 1,048,576 tokens, so you can pass a large document or an entire set of files in one request and work with them as a whole.
Yes. The model data indicates support for reasoning before answering: particle-1 can break down a task step-by-step before providing a result. This is useful for logic, mathematics, code analysis, and situations where a verified conclusion is more important than the first draft.
Yes, both features are supported. particle-1 supports function calling, meaning it can interact with external tools and APIs according to your schema, as well as structured JSON output — convenient when the response needs to be passed directly into code or a system without manual parsing.
Of the non-text formats, particle-1 accepts images and documents — they can be submitted along with the task text. Audio and video support is not indicated in our data, so do not rely on it: for such files, it is better to prepare a text transcript in advance.
There is no separate quality assessment for Russian in our data, so we cannot promise a specific level. The model analyzes the task and responds based on your phrasing, and the easiest way to check text quality is with your own examples — for instance, using standard documents or code snippets.