by ~openaiReleased 27 April 2026
A faster, more efficient GPT-5.4 for production at scale, serving chat apps, coding assistants and agent workflows with reliable instruction following and solid multi-step reasoning over text and images
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 | 249,63 DA | 1 497,76 DA |
| Cached tokens read | 24,96 DA |
per 1M tokens
Web search · 3,33 DA per call
One company runs this model. OpenDunes sends every request to it. DA per 1M tokens: ↑ what you send · ↓ what comes back.
| Provider | ↑ Sent | ↓ Back | Latency | Throughput | Uptime |
|---|---|---|---|---|---|
| OpenAI: GPT Mini Latest | 249,63 DA | 1 497,76 DA | … | … | … |
Also charged: web search 3,33 DA per search.
You pay the price of the provider that served your request. 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="~openai/gpt-mini-latest",
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_tokensseedReasoning
Let it think before answering — and see the thinking.
reasoningreasoning_effortinclude_reasoningStructured output
Get JSON you can parse instead of prose.
response_formatstructured_outputsTools
Let the model call your functions.
toolstool_choiceOther
Also accepted by this model.
max_completion_tokensverbosityAbout 1,00 DA per message. ↑ 249,63 DA · ↓ 1 497,76 DA per 1M tokens.
No. You pay per use from your balance, with no subscription. Try it in Chat at the same price.
A faster, more efficient GPT-5.4 for production at scale, serving chat apps, coding assistants and agent workflows with reliable instruction following and solid multi-step reasoning over text and images.
Deposit dinars with your Edahabia or CIB card. Every model, GPT Mini Latest included, is paid from that balance. No foreign card, no conversion.
~openai names no languages for GPT Mini Latest. Arabic, French and Darja: not stated.
It reads text, images and PDFs. Up to 400K tokens in one request, about 400 pages. It writes up to 96 000 words at once.
No. ~openai has not published its weights.
~openai. Released 27 April 2026. It knows the world up to 31 August 2025.
#2 for reasoning · GPT-6 Astra reads your text and pictures and writes back, taking in about a thousand pages at once and writing up to 128 pages in one answer. It is the flagship of the GPT-6 series, aimed at big end-to-end jobs such as analysis, building software, researching a subject in depth, science and writing documents, with extra strength on long tasks that drive a computer or browser. GPT-6 Sol is the cost-efficient step below, GPT-6 Luna the fast one, and GPT-6 Astra Pro adds a deeper-thinking mode.
GPT-6 Luna reads your text, pictures and files and writes back, and it can take in about a thousand pages at once. It is the quick, low-cost tier of the GPT-6 series, below GPT-6 Sol, built for heavy traffic and tight response times: chat, sorting messages into categories and light multi-step tasks. Turn its thinking up and it can also take software and computer-use jobs that used to need a Sol-class model.
#3 for reasoning · GPT-6.1 Sol reads your text, pictures and files and writes back, and it can take in about a thousand pages at once. It upgrades GPT-6 Sol and sits below the flagship Astra, aimed at coding on its own over many steps, operating a computer, professional work built on large documents and multi-step business automation. Compared with the model it replaces, it gets facts wrong less often and follows stated limits and the user's intent more closely, while landing near Astra on these jobs for far less.
#5 for reasoning · This model reads text, pictures and files, taking in about a thousand pages at once, and writes back as many as 120 pages. It thinks before answering, can search the web and run the tools you give it, so it suits everyday coding, writing and multi-step work at the computer. It stands out at making sense of screenshots, and it sits in the middle of its family: the bigger relation above it reads the same amount but costs more to run. It gives back only text, so a job that needs a picture or a voice reply calls for a different kind of model.
From what its maker and two benchmark sites publish, and where it stands on OpenDunes. Read 7 Nov 2026.
Thinks first
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
Can look things up online while it answers.
Cheaper on repeats
A prompt it saw a moment ago costs less the second time.
Open weights No.
Read on developers.openai.com.
Knows events up to 31 Aug 2025
What it learned stops at this date.
Read a long PDF contract and list what to check
Files
Run an agent that calls your tools
Uses your tools · Thinks first
Explain a screenshot of an error
Images