by DeepSeekReleased 17 October 2026
#10 for 3D scenesThis is the 17 October 2026 snapshot of V4.1 Flash, the image-and-text model DeepSeek serves under its current Flash name. It keeps about a thousand pages in mind, replies up to about 393 pages, thinks first, uses your connected tools and fills in the exact form you give it. Pick it for agents and coding work that include screenshots, and when you want the newest Flash. The earlier DeepSeek V4.1 Flash is the snapshot before this one; V4 Pro 0813 is the larger model for hard reasoning.
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 | 49,93 DA | 199,70 DA |
| Cached tokens read | 1,00 DA |
per 1M tokens
40 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 |
|---|---|---|---|---|---|---|
| Relace | 2,00 DA | 166,42 DA | 1,66 DA | … | … | … |
| OpenInference | 5,82 DA | 199,70 DA | 1,33 DA | … | … | … |
| MorphUnited States | 7,32 DA | 133,13 DA | 1,33 DA | … | … | … |
| WaferUnited States | 12,65 DA | 166,42 DA | 3,33 DA | … | … | … |
| Particle AIUnited States | 16,64 DA | 166,42 DA | 1,00 DA | … | … | … |
| Wafer | 21,97 DA | 166,42 DA | 2,16 DA | … | … | … |
| AlibabaAsia | 24,96 DA | 99,85 DA | 2,50 DA | … | … | … |
| InferenceNetUnited States | 26,63 DA | 183,06 DA | 2,66 DA | … | … | … |
| InferenceNet | 26,63 DA | 183,06 DA | 2,66 DA | … | … | … |
| Baidu65 % off | … | … | … | |||
| DeepInfra51 % offUnited States | … | … | … | |||
| Cloudflare | 33,28 DA | 199,70 DA | 1,00 DA | … | … | … |
| Sail ResearchUnited States | 33,62 DA | 49,93 DA | 2,24 DA | … | … | … |
| PrimeIntellect | 36,35 DA | 399,40 DA | 2,00 DA | … | … | … |
| StreamLake63 % off | … | … | … | |||
| Decart18 % offUnited States | … | … | … | |||
| StepFun55 % off | … | … | … | |||
| DeepSeek | 49,93 DA | 199,70 DA | 1,00 DA | … | … | … |
| Makora | 49,93 DA | 399,40 DA | 2,00 DA | … | … | … |
| Sail Research | 50,42 DA | 74,89 DA | 3,36 DA | … | … | … |
| Reka | 53,25 DA | 366,12 DA | 2,66 DA | … | … | … |
| IonstreamUnited States | 54,92 DA | 164,75 DA | 3,49 DA | … | … | … |
| Novita45 % off | … | … | … | |||
| ModelRunUnited States | 56,58 DA | 223,00 DA | 2,00 DA | … | … | … |
| DekaLLMAsia | 59,91 DA | 226,33 DA | 2,00 DA | … | … | … |
| DigitalOcean | 59,91 DA | 239,64 DA | 2,00 DA | … | … | … |
| GMICloud40 % offUnited States | … | … | … | |||
| Cohere | 64,90 DA | 219,67 DA | 2,00 DA | … | … | … |
| AtlasCloud | 64,90 DA | 259,61 DA | 1,30 DA | … | … | … |
| Phala | 66,57 DA | 399,40 DA | 2,00 DA | … | … | … |
| Fireworks · Spot | 69,90 DA | 279,58 DA | 1,40 DA | … | … | … |
| PrimaLabs | 79,88 DA | 319,52 DA | 1,60 DA | … | … | … |
| Modal | 83,21 DA | 332,84 DA | 1,66 DA | … | … | … |
| Crusoe | 96,52 DA | 399,40 DA | 2,33 DA | … | … | … |
| SiliconFlowUnited States | 99,85 DA | 183,06 DA | 2,00 DA | … | … | … |
| CoreWeaveUnited States | 99,85 DA | 399,40 DA | 2,00 DA | … | … | … |
| Near AI | 99,85 DA | 399,40 DA | 2,00 DA | … | … | … |
| VeniceUnited States | 99,85 DA | 399,40 DA | 2,50 DA | … | … | … |
| Mistral | 99,85 DA | 399,40 DA | 2,33 DA | … | … | … |
| Nebius | 99,85 DA | 399,40 DA | 2,00 DA | … | … | … |
| Together | 99,85 DA | 399,40 DA | 2,00 DA | … | … | … |
| Fireworks | 99,85 DA | 399,40 DA | 2,00 DA | … | … | … |
| BaseTen | 99,85 DA | 399,40 DA | 2,33 DA | … | … | … |
| Parasail | 99,85 DA | 399,40 DA | 2,00 DA | … | … | … |
| FireworksUnited States | 149,78 DA | 599,10 DA | 3,00 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="deepseek/deepseek-v4.1-flash-1017",
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_choiceparallel_tool_callsOther
Also accepted by this model.
max_completion_tokensAbout 0,15 DA per message. ↑ 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.
#10 for 3D scenes · This is the 17 October 2026 snapshot of V4.1 Flash, the image-and-text model DeepSeek serves under its current Flash name. It keeps about a thousand pages in mind, replies up to about 393 pages, thinks first, uses your connected tools and fills in the exact form you give it. Pick it for agents and coding work that include screenshots, and when you want the newest Flash. The earlier DeepSeek V4.1 Flash is the snapshot before this one; V4 Pro 0813 is the larger model for hard reasoning.
DeepSeek V4.1 Flash costs 0,040 DA less for the same message.
Deposit dinars with your Edahabia or CIB card. Every model, DeepSeek V4.1 Flash 1017 included, is paid from that balance. No foreign card, no conversion.
DeepSeek names no languages for DeepSeek V4.1 Flash 1017. Arabic, French and Darja: not stated.
It reads text and images. Up to 1M tokens in one request, about 1 000 pages. It writes up to 295 000 words at once.
Yes, its weights are public. Open weights on Hugging Face
DeepSeek. Released 17 October 2026.
DeepSeek V4.1 Flash is the first DeepSeek model built on a new design, and it reads images as well as text, keeping about a thousand pages in mind. It thinks before answering, uses your connected tools and fills in the exact form you give it. Use it as the fast, everyday model for coding agents that also need to look at screenshots or diagrams.
DeepSeek's generally available V4 Flash, re-post-trained for coding, reasoning and agent work on the same 284B mixture-of-experts with 13B active, with tools, structured output and switchable reasoning
DeepSeek V4 Flash is the first V4 release of the Flash size, built for fast answers over very long text. It reads about a thousand pages at once, thinks before answering when asked, uses your connected tools and fills in the exact form or table you give it. Pick it to run an agent or summarise a whole codebase quickly. DeepSeek V4 Flash 0731 is the retrained revision for coding and agents; V4 Pro is the larger model.
DeepSeek V4 Pro 0813 is the general-availability release of V4 Pro, the large model for hard reasoning, coding and long agent runs. It reads about a thousand pages of text, replies up to about 384 pages, uses your connected tools and fills in the exact form you give it, with three thinking levels (low, high and max) so you choose how much it reasons.
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.
Read image-and-text dumps up to a million tokens and reason over them in one pass
Thinks first
Drive agent loops that call your tools across long multi-step tasks
Uses your tools · Thinks first
Rerun near-identical large prompts at the cached-input rate
Cheaper on repeats