by MiniMaxReleased 19 October 2026
This model accepts text, pictures, and video inputs, and can process up to 1000 pages of information to generate responses up to 520 pages long. It excels at complex tasks like coding, cybersecurity, and advanced problem-solving. Its ability to handle video input and its high ranking in reasoning set it apart from other models. However, it cannot search the web or structure data into specific formats.
You pay for what each request uses. No subscription, no minimum.
50 % offSambaNova has cut its price. Every figure here is the new one.
One message, the ordinary use
Every rate
| Meter | ↑ you send | ↓ comes back |
|---|---|---|
| Text tokens | 49,93 DA | 199,70 DA |
| Cached tokens read | 4,99 DA |
per 1M tokens
Twelve 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 | Cache read | Latency | Throughput | Uptime |
|---|---|---|---|---|---|---|
| GMICloud25 % offUnited States | … | … | … | |||
| StreamLake10 % off | … | … | … | |||
| Fireworks | 49,93 DA | 199,70 DA | 4,99 DA | … | … | … |
| InferenceNet | 49,93 DA | 199,70 DA | 4,99 DA | … | … | … |
| DigitalOcean | 49,93 DA | 199,70 DA | 4,99 DA | … | … | … |
| Novita | 49,93 DA | 199,70 DA | 4,99 DA | … | … | … |
| Nebius | 49,93 DA | 199,70 DA | 4,99 DA | … | … | … |
| SambaNova50 % off | … | … | … | |||
| ModelRunUnited States | 49,93 DA | 199,70 DA | 4,99 DA | … | … | … |
| Parasail | 49,93 DA | 199,70 DA | 4,99 DA | … | … | … |
| InferenceNetUnited States | 49,93 DA | 199,70 DA | 4,99 DA | … | … | … |
| Together | 49,93 DA | 199,70 DA | 4,99 DA | … | … | … |
| MinimaxUnited States | 49,93 DA | 199,70 DA | 4,99 DA | … | … | … |
You pay the price of the provider that served your request. Prices read 15:06 · 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="minimax/minimax-m3.1-flash",
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_logprobsReasoning
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_choiceAbout 0,15 DA per message. Was 0,30 DA. ↑ 49,93 DA · ↓ 199,70 DA per 1M tokens.
No. You pay per use from your balance, with no subscription. Try it in Chat at the same price.
This model accepts text, pictures, and video inputs, and can process up to 1000 pages of information to generate responses up to 520 pages long. It excels at complex tasks like coding, cybersecurity, and advanced problem-solving. Its ability to handle video input and its high ranking in reasoning set it apart from other models. However, it cannot search the web or structure data into specific formats.
Deposit dinars with your Edahabia or CIB card. Every model, MiniMax M3.1 Flash included, is paid from that balance. No foreign card, no conversion.
MiniMax names no languages for MiniMax M3.1 Flash. Arabic, French and Darja: not stated.
It reads text, images and video. Up to 1M tokens in one request, about 1 000 pages. It writes up to 393 000 words at once.
Yes, its weights are public. Open weights on Hugging Face
MiniMax. Released 19 October 2026.
MiniMax H3 edits and generates video from text, pictures, video and audio, with 2K output, built-in sound and clips of 5 to 15 seconds. It follows written edit instructions, shows text and logos legibly and copies motion from one video to another, which suits ads, product videos and game or app mockups. You can fix the first and last frames.
MiniMax Speech 2.8 HD turns text into speech and returns audio, using any MiniMax voice you name. It is the quality tier: it voices laughs and sighs written into the text as tags, and it supports emotion settings like happy, sad or calm. Use it for narration, audiobooks and finished voice-overs.
MiniMax's Hailuo 2.3, aimed at cinematic scenes and character animation where motion must look real and characters expressive, producing 6 or 10 second 1080p clips from a prompt or an opening image
MiniMax H3 Max makes video from a written description or a picture at a lower resolution than H3, in clips of 5 to 15 seconds, and lets you fix the first frame, the last frame or both. It is the faster model, built on MiniMax H3 with extra training, so it suits drafts and quick tries on a shot. Its output is silent.
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 No.
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
The model's weights are public.
Solving intricate problems by thinking through steps first
Thinks first
Executing tasks that require the use of external tools
Uses your tools
Handling repeated requests more cheaply over time
Cheaper on repeats