by GoogleReleased 13 March 2025
Gemma 3 4B is the smallest Gemma 3 model you can download and run yourself: it reads text and images, answers in text, keeps about 128 pages in mind at once and speaks 140+ languages. Unlike the 12B and 27B, it cannot use tools. Pick it for light jobs such as short chat and simple image questions.
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 | 16,64 DA | 33,28 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.
You pay the price of the provider that served your request. Prices read 15:19 · 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="google/gemma-3-4b-it",
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_biaslogprobstop_logprobsStructured output
Get JSON you can parse instead of prose.
response_formatstructured_outputsAbout 0,033 DA per message. ↑ 16,64 DA · ↓ 33,28 DA per 1M tokens.
No. You pay per use from your balance, with no subscription. Try it in Chat at the same price.
Gemma 3 4B is the smallest Gemma 3 model you can download and run yourself: it reads text and images, answers in text, keeps about 128 pages in mind at once and speaks 140+ languages. Unlike the 12B and 27B, it cannot use tools. Pick it for light jobs such as short chat and simple image questions.
Gemma 3 12B costs 0,0090 DA more for the same message.
Deposit dinars with your Edahabia or CIB card. Every model, Gemma 3 4B included, is paid from that balance. No foreign card, no conversion.
Google says Gemma 3 has multilingual support in over 140 languages. Arabic, French and Darja are not named.
It reads text and images. Up to 131K tokens in one request, about 130 pages. It writes up to 12 000 words at once.
Yes, its weights are public. Open weights on Hugging Face
Google. Released 13 March 2025. It knows the world up to 31 August 2024.
This model is designed to handle text, pictures, and video inputs, allowing it to read up to 130 pages of content. It specializes in generating and editing video clips, producing outputs that include both text and moving images. What sets this model apart is its native ability to process and generate video content directly. It is not suitable for tasks requiring external tool usage or the creation of structured data formats.
#1 for speech · Gemini 3.8 Flash TTS turns text into speech in 130+ languages with studio-grade acting, regional accents and a voice that stays consistent through long dialogues. It offers prebuilt voices, custom voice personas and voice replication, with about 8 pages of text per request. Pick it for audiobooks, narration and multi-speaker scenes.
Nano Banana 2, Google's 3.1 Flash Image model, generates and edits images from a written description and reference pictures, returning images with text. It outputs up to 4K, adds very tall and very wide shapes such as 1:8, and renders text in other languages better. Pick it for production batches of ads, banners and edits.
This model accepts text, pictures, and videos as input, allowing it to read up to 130 pages of content. It specializes in creating and editing video content, converting descriptions or pictures into moving images. A key feature is its ability to search the web for information while processing complex visual data. However, it does not utilize external tools or generate structured data formats.
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 No.
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 product photos
Images
Return a message's key fields as JSON
Fills your JSON
Translate short messages