Llama 4 Maverick
Neural Network
Llama 4 Maverick is Meta's multimodal mixture-of-experts model for text and image processing
Specifications
- Input
- text, images, documents
- Output
- text
- Context size
- 524 288 tokens
- Max answer length
- 32 768 tokens
- Capabilities
- 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: 'llama-4-maverick',
});
}
// Async - Text generation
async function main() {
const stream = await openai.chat.completions.create({
messages: [{ role: 'user', content: 'Say this is a test' }],
model: 'llama-4-maverick',
stream: true
});
for await (const chunk of stream) {
const part: string | null = chunk.choices[0].delta?.content ?? null;
}
}
main();
How it works Llama 4 Maverick?
Llama 4 Maverick is a multimodal model from Meta based on the mixture-of-experts architecture: out of 128 experts, only 17 billion parameters are activated per pass, combining large architecture capacity with efficient computation. The Instruct version is fine-tuned for instructions and dialogue, and multimodality allows inputting both text and images. In BotHub, you can use it without a VPN or foreign card: pay with Russian cards in rubles based on usage; tokens do not expire. Over 250 neural networks are available in one window, making it easy to compare Llama 4 Maverick with other models, while a unified OpenAI-compatible API allows switching between them without rewriting integrations. Requests and responses are encrypted; for companies, we offer contracts, invoices, EDI, and an admin panel with limits. It is useful for analyzing screenshots or diagrams, building an assistant that processes text and images, breaking down long technical descriptions into clear steps, assisting with code and debugging, or running the same query through multiple models to choose the best answer.Frequently asked questions about Llama 4 Maverick
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.
Llama 4 Maverick works with text: writes and edits, analyzes code, answers based on documents, and analyzes images. A context window of up to 524,288 tokens allows uploading large files and entire long conversations. It supports function calling, making it easy to integrate into services and automations.
The model generates up to 16,384 tokens per response — enough for a long article, detailed analysis, or a large code snippet. If you need more, ask it to continue from where it stopped or break the task into parts: the 524,288-token context will retain all previous text.
A specific reasoning mode is not noted in our data, so we won't promise it. In practice, a simple trick helps: ask in your prompt to break down the task step-by-step, list the conditions, and only then provide a conclusion — answers involving logic and calculations become more accurate.
Yes, function calling and structured JSON output are supported, so the model can be connected to search, databases, or internal services to receive answers in a specified schema. In BotHub, this works via a unified OpenAI-compatible API — you can switch models without rewriting integrations.
Images and documents — yes: you can send a screenshot, diagram, photo, or file and ask it to analyze the content. Audio and video are not listed as supported input formats; for such tasks, BotHub has separate models available in the same window.
Working with Russian is not separately described in our data, so we recommend testing the model with your own task — a couple of test prompts will show more than any promises. On the BotHub side, everything is in Russian: the interface, support, and payment with Russian cards, without VPN or foreign cards.