DeepSeek models
All 16 DeepSeek models on OpenDunes, priced in Algerian Dinars.
16 models
#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 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.
DeepSeek V4 Pro is the large V4 model, reading about a thousand pages of text and replying up to about 384 pages. It is for the hardest reasoning and coding jobs and for agents that run for many steps, using your connected tools and filling in the exact form you give it. DeepSeek V4 Pro 0813 is the general-availability release of this model.
DeepSeek V3.2 reads long inputs efficiently, and it is tuned for reasoning and for using connected tools inside agents. It takes text and answers in text, keeping about 160 pages in mind with replies up to about 65 pages, and you can switch thinking on or off. Use it for agents that plan and use many tools over a long document. DeepSeek V4 Flash reads about a thousand pages; V3.2 Exp is its experimental predecessor.
DeepSeek V4 Flash Vision Exp adds images to V4 Flash 0731: you send text with screenshots, charts or photos and it answers in text, still keeping about a thousand pages in mind. On text, tool use and agent tasks it is meant to match the version it is built on. It is marked experimental, so use it to try image reading in an agent or a document workflow.
DeepSeek V3.1 is a hybrid model: the same network answers directly or thinks first, and you choose per request. It reads about 160 pages of text, uses connected tools more reliably than V3, and fills in the exact form or table you give it. Use it when one model must serve both quick chat and harder reasoning, for example an agent that sometimes needs to plan. DeepSeek V3.1 Terminus fixes its language mixing; DeepSeek V3.2 is the later generation.
DeepSeek V3 0324 is a chat model that improved on the first V3 in reasoning, frontend web code, Chinese writing and using connected tools. It takes text, answers in text, keeps about 160 pages in mind and can reply up to about 115 pages. Pick it for web page generation, long drafts and tool-using bots that do not need a visible thinking step.
DeepSeek V3 is the plain chat model of the V3 line that does not think first: text in, text out, keeping about 160 pages in mind with replies up to about 16 pages, and it can use connected tools and fill in exact forms. Use it for drafting, summaries, translation and everyday coding help where you want an answer straight away. Newer releases have replaced it.
DeepSeek V3.2 Exp is the experimental release that first tried the efficient way of reading long text found in V3.2, between V3.1 and V3.2. You send text and it answers in text, keeping about 160 pages in mind and replying up to about 147 pages; thinking is optional, and it can use connected tools and fill in exact forms. Use it only to test that design or reproduce older results.
DeepSeek R1 0528 is the May 2025 update of R1, a thinking model that shows its reasoning before the answer. It scores higher on maths and science tests, invents fewer facts, and adds the ability to fill in an exact form and to use connected tools. It reads about 160 pages of text and replies up to about 32 pages, so use it for hard maths, logic and code debugging.
DeepSeek V3.1 Terminus is a maintenance release of V3.1 that fixes replies mixing languages and strengthens agents that write code and agents that search. It keeps the choice between thinking and answering directly, reads about 160 pages of text and allows replies up to about 147 pages. Use it for agents that browse, search and edit code, and for multilingual answers that must stay in one language. DeepSeek V3.2 is the next generation.
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