by MetaReleased 5 April 2025
Llama 4 Scout takes text and pictures and answers in text or code across 12 languages. It reads about 1,300 pages at once, enough to caption or question many images, or take in a very large document in one request. Meta tuned it for chat in many languages, captions and image understanding.
You pay for what each request uses. No subscription, no minimum.
One message, the ordinary use
Every rate
| Meter | ↑ you send | ↓ comes back |
|---|---|---|
| Text tokens | 33,28 DA | 99,85 DA |
per 1M tokens
Two companies run this model. OpenDunes sends each request to one of them, picked by price and speed. If it fails, the next one takes it. DA per 1M tokens: ↑ what you send · ↓ what comes back.
| Provider | ↑ Sent | ↓ Back | Latency | Throughput | Uptime |
|---|---|---|---|---|---|
| DeepInfraUnited States | 33,28 DA | 99,85 DA | … | … | … |
| GoogleUnited States | 83,21 DA | 232,98 DA | … | … | … |
You pay the price of the provider that served your request. Prices read 15:20 · speed over the last 30 minutes · uptime over 3 days.
Real requests, real replies. Try any of them yourself — one click, pre-filled.
What's the total on this invoice, and when is it due?
invoice-scan.jpg · photoThe total is 42 300,00 DA, due 5 September 2026. Worth a double-check: the handwritten day could read as 8 instead of 3.
Try in Chat LabTurn this customer email into JSON with intent, urgency and order_id.
customer-email.txtTry your own request — same model, pre-selected.
Open Chat LabSwap the base URL and the key. Your OpenAI code runs unchanged.
import os
from openai import OpenAI
client = OpenAI(
base_url="https://opendunes.com/v1",
api_key=os.environ["OPENDUNES_API_KEY"],
)
resp = client.chat.completions.create(
model="meta-llama/llama-4-scout",
messages=[{"role": "user", "content": "Two lines to welcome a shop in Oran."}],
)
print(resp.choices[0].message.content)Runs this request once on your balance.
Standard OpenAI fields, grouped by what they control. Anything not listed is accepted and ignored.
Shaping the reply
How long, how varied, where to stop.
max_tokenstemperaturetop_ptop_kmin_pstopseedfrequency_penaltypresence_penaltyrepetition_penaltylogit_biasStructured output
Get JSON you can parse instead of prose.
response_formatstructured_outputsTools
Let the model call your functions.
toolstool_choiceAbout 0,083 DA per message. ↑ 33,28 DA · ↓ 99,85 DA per 1M tokens.
No. You pay per use from your balance, with no subscription. Try it in Chat at the same price.
Llama 4 Scout takes text and pictures and answers in text or code across 12 languages. It reads about 1,300 pages at once, enough to caption or question many images, or take in a very large document in one request. Meta tuned it for chat in many languages, captions and image understanding.
Llama 4 Maverick costs 0,20 DA more for the same message.
Deposit dinars with your Edahabia or CIB card. Every model, Llama 4 Scout included, is paid from that balance. No foreign card, no conversion.
Meta supports 12 languages: Arabic, English, French, German, Hindi, Indonesian, Italian, Portuguese, Spanish, Tagalog, Thai and Vietnamese. Darja is not named.
It reads text and images. Up to 1.3M tokens in one request, about 1 300 pages. It writes up to 12 000 words at once.
Yes, its weights are public. Open weights on Hugging Face
Meta. Released 5 April 2025. It knows the world up to 31 August 2024.
Llama 3.1 8B Instruct is the small model of Meta's Llama 3.1 family, tuned to follow instructions and described as fast and efficient. You send text and get text back; it reads about 131 pages at once and can use tools or fill in an exact form you give it. Use it for sorting, short rewrites, simple chat and high-volume jobs.
Llama 3.3 70B Instruct is a text-only Meta model tuned for conversation in English, Spanish, French, German, Italian, Portuguese, Hindi and Thai. You send a chat, with tools or an exact form to fill in if you need them, and get a reply of up to about 16 pages from a conversation of up to 131 pages. Pick it for assistants and support in those languages.
Llama 4 Maverick takes text and pictures and answers in text or code across 12 languages, keeping about a thousand pages of conversation in mind. Meta tuned it for reasoning about images and for assistant-style chat. Use it to answer questions about photos, charts and screenshots, or for multilingual assistants with long conversations.
Llama Guard 4 checks whether a request or a model's reply is safe. You send text, or text with pictures, and it answers in text: safe, or unsafe with the violated categories from a published list of hazards. Use it as a filter before a chat model sees a user message, and again on the reply. It sorts and does not chat or answer questions, so pair it with a chat model such as Llama 4 Scout.
From what its maker and two benchmark sites publish, and where it stands on OpenDunes. Read 7 Nov 2026.
Thinks first No.
Works through the problem before it answers.
Uses your tools
Can call functions your code provides.
Fills your JSON
Answers in the JSON shape you give it.
Searches the web No.
Can look things up online while it answers.
Cheaper on repeats No.
A prompt it saw a moment ago costs less the second time.
Open weights
The model's weights are public.
Knows events up to 31 Aug 2024
What it learned stops at this date.
Caption a batch of product photos
Images
Answer support questions in Arabic and French
Search a long log file for the first error