Structured output
You send
Turn this customer email into JSON with intent, urgency and order_id.
customer-email.txtMercury 3
Newby InceptionReleased 16 October 2026
Mercury 3 is a diffusion-based text model that thinks first and calls your tools, built for agentic workflows and real-time work where latency compounds, with a million-token window and JSON output
TakesText
GivesText
This model is being retired on Nov 8, 2026. Plan a migration before then.
What it costs
You pay for what each request uses. No subscription, no minimum.
≈ 0,092 DA per message3 000 DA buys about 32 000 messages
50 % offInception has cut its price. Every figure here is the new one.
One message, the ordinary use
1 000 tokens sent× 33,28 DA per 1M0,033 DA
500 tokens back× 116,49 DA per 1M0,058 DA
Total≈ 0,092 DA
Every rate
| Meter | ↑ you send | ↓ comes back |
|---|---|---|
| Text tokens | 33,28 DA | 116,49 DA |
| Cached tokens read | 3,33 DA |
per 1M tokens
Running a tool-using agent loop that plans, calls your functions and returns structured results≈ 0,27 DA
Filling a strict JSON schema from messy customer tickets in batches≈ 0,40 DA
Reasoning through a hard multi-step problem with extended thinking turned up≈ 0,51 DA
Drafting and revising long reports against a large knowledge base pasted in one go≈ 3,93 DA
Answering questions from a full codebase dropped into the context in a single pass≈ 6,73 DA
- When you pay
- After the answer, from your balance. Nothing is held.
- Why it varies
- A longer question or a longer answer costs more.
- After each one
- The exact charge shows on the request.
DepositDeposits from 50 DA, with Edahabia or CIB.
Where your request runs
One company runs this model. OpenDunes sends every request to it. DA per 1M tokens: ↑ what you send · ↓ what comes back.
Inception50 % off…
↑ Sent, per 1M
↓ Back, per 1M
Cache read
Latency
…
Throughput
…
You pay the price of the provider that served your request. Prices read 15:18 · speed over the last 30 minutes · uptime over 3 days.
See it in action
Real requests, real replies. Try any of them yourself — one click, pre-filled.
Writing
You send
Two lines to welcome a shop in Oran.
It replied
Welcome to our new shop in Oran — fresh bread every morning, and coffee the way the city likes it. Come in, the first cup is on us.
Try in Chat LabTry your own request — same model, pre-selected.
Open Chat LabCall it in code
Swap 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="inception/mercury-3",
messages=[{"role": "user", "content": "Two lines to welcome a shop in Oran."}],
)
print(resp.choices[0].message.content)pip install openaiFull reference for this route
Runs this request once on your balance.
Parameters you can send
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_tokenstemperaturestopReasoning
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_choiceQuestions about Mercury 3
How much does Mercury 3 cost in dinars?
About 0,092 DA per message.
Was 0,18 DA. ↑ 33,28 DA · ↓ 116,49 DA per 1M tokens.Is Mercury 3 free? Can I try it?
No. You pay per use from your balance, with no subscription. Try it in Chat at the same price.
What is Mercury 3 best for?
Mercury 3 is a diffusion-based text model that thinks first and calls your tools, built for agentic workflows and real-time work where latency compounds, with a million-token window and JSON output.
How do I pay for Mercury 3 without an international card?
Deposit dinars with your Edahabia or CIB card. Every model, Mercury 3 included, is paid from that balance. No foreign card, no conversion.
Does Mercury 3 work in Arabic, French or Darja?
Inception names no languages for Mercury 3. Arabic, French and Darja: not stated.
Can it read PDFs, images or a whole book?
It reads text. Up to 1M tokens in one request, about 1 000 pages. It writes up to 98 000 words at once.
Is Mercury 3 open source?
No. Inception has not published its weights.
Who makes Mercury 3, and when was it released?
Inception. Released 16 October 2026.
Other Inception models, priced in dinars
See all Inception modelsMercury DecideInception · Sep 30, 2026
Mercury Decide takes a situation written as text, with pictures if you like, plus questions of a fixed kind. For each question it gives back a yes or no, a pick from options, or a numeric score, with a probability worked out inside the model and no written answer. It handles up to 14 decisions a second, so you can run every case through it and send doubtful ones to a person. Kev 4B is the alternative you can download and run yourself, for situations that are text only.
65 pages per call
≈0,0067 DAper case
View modelMercury 2.5Inception · Sep 8, 2026
Mercury 2.5 writes by refining a whole stretch of text at once instead of word by word, reaching more than a page of text every second, and it thinks at a level you set. You send up to 260 pages of text, with tools or an exact form to fill in, and it can use several tools at the same time. Use it in coding helpers, search helpers and voice systems, where each model call adds delay.
Thinks firstUses your tools260 pages at once
≈0,038 DAper message
View modelMercury 2Inception · Mar 4, 2026
Mercury 2 thinks before it answers and writes by refining a whole stretch of text at once instead of word by word, so it produces over a page of text every second. You send up to 128 pages of text, and it can use your tools or fill in an exact form you give it. Use it where waiting adds up: voice assistants, live search and coding helpers that call the model many times.
Thinks firstUses your tools120 pages at once
≈0,21 DAper message
View model