DeepSeek V3.1
DeepSeek · released Aug 21, 2025 · deepseek-ai/DeepSeek-V3.1
- Type
- Open weightsMIT License
- Params
- 685B
- Context
- 164K
37B active per word · about 123K words of context
Our take
Written Sep 30, 2026DeepSeek V3.1 is a text model you can download and run yourself, or reach through a host, under terms that allow commercial use, changes and redistribution. Its measured quality is mid-field rather than leading, so it is a budget pick for everyday work rather than a model to choose for the hardest tasks.
Use it for everyday text generation and chat where the bill matters more than peak quality, or when you want a large model you can run yourself with a licence that allows commercial use, changes and redistribution. Its request capacity takes long documents without splitting them up first, though reliable recall across all of it is unverified in our data. Skip it if a task needs top-tier measured coding or reasoning, or if you meant to run it on a machine that cannot hold a model this size.
The case for it
- The licence allows commercial use, changes and redistribution (MIT), so the terms are not the thing that decides whether you can build on it.
- A request capacity of 163,840 tokens means long documents need not be split up first, though room to hold them is not a guarantee of accurate recall.
- Ten hosted offers, so a rate rise at one host is a reason to move the workload rather than a bill to absorb.
The case against it
- Mid-field on the boards we hold: 80th of 168 on Arena Text (overall) as of 25 Sep 2026, and 96th of 168 on Arena Coding as of 25 Sep 2026.
- At 685 billion parameters in total, with 37 billion active per token, running it yourself needs hardware well beyond a workstation.
How good is it?
EverydayGeneral questions and everyday reasoning
Arena Text (overall)80th of 168 · 1417
Also on this board: 1416 (Sep 25, 2026). Read the pair, not the higher one.
CodingWriting and fixing code on its own
Arena Coding96th of 168 · 1447
Arena Coding is the only board that has scored it for this.
Also on this board: 1455 (Sep 25, 2026). Read the pair, not the higher one.
AgenticPlanning, calling tools, staying on task
Not yet scored on Arena Agent.
WritingDrafting and rewriting prose
Arena Creative Writing81st of 168 · 1388
Arena Creative Writing is the only board that has scored it for this.
Also on this board: 1402 (Sep 25, 2026). Read the pair, not the higher one.
These tests check whether a model follows instructions — a precondition for all the work above, but not a measure of how well that work is done.
Every published score for this model6 scoresEvery figure we hold, from 6 boards, with who ran it and a link to the source — including the boards no rating above is built on.
Can you run it yourself?
GeForce RTX 4090 · 24 GB
Too large for this card. The weights do not fit even with part of them offloaded to system memory.
Apple M1 Pro (16-core GPU) · 32 GB
Too large for this card. The weights do not fit even with part of them offloaded to system memory.
Radeon RX 7900 XT · 20 GB
Too large for this card. The weights do not fit even with part of them offloaded to system memory.
Memory use by level
Against a 24 GB card.
What is quantisation? →This model on every device we track71 devicesThe Q4 build most people download, on each device: what the weights come to, how much context the memory leaves, and whether it runs. Smallest device that runs it first.
Check against your own machine → · Where to rent it hosted →
Or rent it from someone else
Prices checked between 60 min and 4 days ago — each listing carries its own date.
Some hosts sell this model at two prices: on their own price list (“direct”) and on their OpenRouter listing (“through OpenRouter”). Where the two differ, the row shows both, each with the date we last read it.
- per 1M tokens
- $0.25 in / $0.95 out
- Context served
- 164K
- Throughput
- Not measured
| Provider | In / out per 1M tokens | Context | Throughput | Trains on prompts | Logs prompts | Zero retention |
|---|---|---|---|---|---|---|
| DeepInfrafp4Direct and through OpenRouter | $0.25 / $0.95checked 1 hour ago | 164K33K max reply through OpenRouter | 6 tok/sthrough OpenRouter | DirectUnknownThrough OpenRouterNo | DirectUnknownThrough OpenRouterNo | DirectUnknownThrough OpenRouterConfirmed |
| OpenRouterOpenRouter's own listing | $0.25 / $0.95checked 1 hour ago | 164K | not measured | Unknown | Unknown | Unknown |
| SiliconFlowfp8Through OpenRouter | $0.27 / $1.00checked 1 hour ago | 164K147K max reply | 17 tok/s | No | No | Confirmed |
| Novita AIDirect | $0.27 / $1.00checked 4 days ago | 131K | not measured | Unknown | Unknown | Unknown |
| AtlasCloudfp8Through OpenRouter | $0.30 / $1.00checked 1 hour ago | 131K66K max reply | 44 tok/s | No | Yesunknown period | Unknown |
| SambaNovafp8Direct and through OpenRouter | $3.00 / $4.50directchecked 60 min ago$0.65 / $1.50through OpenRouterchecked 1 hour ago | 131K7K max reply through OpenRouter | 45 tok/sthrough OpenRouter | DirectUnknownThrough OpenRouterNo | DirectUnknownThrough OpenRouterNo | DirectUnknownThrough OpenRouterConfirmed |
| CoreWeavefp8Through OpenRouter | $0.55 / $1.65checked 1 hour ago | 161K145K max reply | 50 tok/s | No | No | Confirmed |
| MaraThrough OpenRouter | $0.60 / $1.70checked 1 hour ago | 131K118K max reply | 94 tok/s | No | No | Confirmed |
DeepInfra
Novita AI
SambaNovaAcross the 8 listings we hold: 6 say they do not train on prompts (2 of them only through OpenRouter), 0 say they do and 2 do not say. 5 appear in the zero-retention registry we check (2 of them only through OpenRouter); the rest are unknown to us.
What each host's API supports
From the parameter list each endpoint publishes. Streaming is omitted: nothing we hold reports it, for any model.
| Provider | Tool calling | JSON output | Strict schema |
|---|---|---|---|
| DeepInfrafp4Direct and through OpenRouter | ✓ | ✓ | ✓ |
| OpenRouterOpenRouter's own listing | ✓ | ✓ | ✓ |
| SiliconFlowfp8Through OpenRouter | ✓ | ✓ | ✓ |
| Novita AIDirect | |||
| AtlasCloudfp8Through OpenRouter | ✗ | ✓ | ✓ |
| SambaNovafp8Direct and through OpenRouter | ✓ | ✗ | ✗ |
| CoreWeavefp8Through OpenRouter | ✓ | ✓ | ✓ |
| MaraThrough OpenRouter | ✗ | ✓ | ✓ |
Tool calling: 5 of 8 listings say yes, 2 say no, 1 publishes no parameter list. JSON output: 6 of 8 listings say yes, 1 says no, 1 publishes no parameter list. Strict schema: 6 of 8 listings say yes, 1 says no, 1 publishes no parameter list.
Models people weigh against DeepSeek V3.1
When we formed this view
Recent changes
What moved
output +5% ($0.95 → $1.00 per 1M tokens), cache read +4% ($0.130 → $0.135 per 1M tokens)What moved
first indexed by our pipelineEach date is the day we first saw the change, or the day the maker announced it.
What we do not know about this model yet
- We hold no measured file for it, so all 3 sizes on this page are calculated from the parameter count.
- 1 of 8 listings publishes no parameter list, so what its API accepts is unknown to us.
- Nothing we hold says whether an endpoint streams, so we do not show it either way.
- 2 of 8 listings do not say whether they train on prompts, and 2 answer only through OpenRouter, not for their own listing.
- We hold no batch or off-peak rate for any of its listings.
- We hold a decode speed for it, but no prompt-processing (prefill) figure, so how long the input side of a job takes is unknown to us.
Licence and identifiers
What the licence allowsMIT License, what it allows commercially, and the identifiers you need to pull this model — its Hugging Face repo, our slug and a machine-readable card.
Licence
MIT License
Fully permissive: do anything with attribution. No patent grant, unlike Apache-2.0.
Identifiers
- Hugging Face
- deepseek-ai/DeepSeek-V3.1
- Architecture
- Mixture of experts
- Takes in, gives back
- Text in, text out
- Catalogue slug
- deepseek-deepseek-v3-1