DeepSeek V3.2 Exp
DeepSeek · released Sep 29, 2025 · deepseek-ai/DeepSeek-V3.2-Exp
- Type
- Open weightsMIT License
- Params
- 685B
- Context
- 164K
37B active per word · about 123K words of context
Our take
Written Jul 28, 2026DeepSeek V3 Experimental is a downloadable text model with a permissive MIT licence and a 163,840-token request limit. It scores well on hard prompts but falls behind on coding and creative writing, with uneven speed across its small set of hosts.
Pick this for budget text inference when you need a permissive licence, or for hard-prompt tasks where its strongest measured score applies. Choose Atlas Cloud or Novita for the highest verified speed. Skip it if you need code generation, creative writing, image or video input, or consistently fast throughput.
The case for it
- Permissive MIT licence with no copyleft, allowing commercial use and redistribution.
- Hard-prompt score 22.9 points above its overall text score on the independent chat leaderboard.
- 685.4 billion total parameters with 37 billion active per token, keeping compute per request moderate.
The case against it
- Code score 153 points below its overall text score on the independent leaderboard.
- Throughput is inconsistent: one host at 12 tokens per second, two of five offers unverified.
- Text-to-text only; no image or video handling.
How good is it?
IntelligencePuzzles, maths, exam questions
Arena Text (overall)59th of 143 · 1422.6
Also on this board: 1424.8 via Thinking (Aug 2, 2026). Read the pair, not the higher one.
CodingWriting and fixing code on its own
Arena Coding60th of 143 · 1464.9
Also on this board: 1475 via Thinking (Aug 2, 2026). Read the pair, not the higher one.
AgenticPlanning, calling tools, staying on task
Nobody we watch has scored DeepSeek V3.2 Exp for this. We would take the rating from Arena Agent (IPS).
WritingWe do not rate this
Two boards come close and neither tests writing: Arena Creative Writing asks people which of two replies they prefer, and LiveBench Language tests whether a model understood a passage. So we show where DeepSeek V3.2 Exp placed and give it no mark out of five.
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, which is why they get no rating.
Every published score for this model7 scoresEvery figure we hold, from 7 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?
- Fits in memory
- weights load entirely on the card
- Spills to system RAM
- some weights offload; much slower
- Too large
- will not load even with offload
- est
- size is calculated; the verdict could change by 10%
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.
All 3 sizes here are calculated, not measured. We hold no measured file for this model, so each size comes from the parameter count and every verdict above inherits that uncertainty.
Check against your own machine → · All 71 devices, with every size →
Or rent it from someone else
Cheapest of 5 live listings. Picked at the widest standard context we hold, within one quantisation slice, so the numbers beside it are a price one host actually charges.
- per 1M tokens
- $0.27 in / $0.41 out
- Context served
- 164K
- Throughput
- Not measured
| Provider | In / out per 1M tokens | Context | Throughput | Trains on prompts | Logs prompts | Zero retention |
|---|---|---|---|---|---|---|
| OpenRouter | $0.27 / $0.41 | 164K | not measured | Unknown | Unknown | Unknown |
| Novita AI | $0.27 / $0.41 | 164K | not measured | Unknown | Unknown | Unknown |
| Novita AIfp8 | $0.27 / $0.41 | 164K | 19 tok/s | No | No | Confirmed |
| SiliconFlowfp8 | $0.27 / $0.41 | 164K | 22 tok/s | No | No | Confirmed |
| AtlasCloudfp8 | $0.27 / $0.41 | 164K | 18 tok/s | No | Yesunknown period | Unknown |
Novita AIAcross the 5 listings we hold: 3 say they do not train on prompts, 0 say they do and 2 do not say. 2 appear in the zero-retention registry we check; the rest are unknown to us rather than confirmed either way.
What each host's API supports
From the parameter list each endpoint publishes. Streaming is omitted: nothing we hold reports it, for any model.
- ✓
- Supported
- ✗
- Not supported
- Not published
- host gave no parameter list
| Provider | Tool calling | JSON output | Strict schema |
|---|---|---|---|
| OpenRouter | ✓ | ✓ | ✓ |
| Novita AI | |||
| Novita AIfp8 | ✓ | ✓ | ✓ |
| SiliconFlowfp8 | ✓ | ✓ | ✓ |
| AtlasCloudfp8 | ✓ | ✓ | ✓ |
Tool calling: 4 of 5 listings say yes, 1 publishes no parameter list. JSON output: 4 of 5 listings say yes, 1 publishes no parameter list. Strict schema: 4 of 5 listings say yes, 1 publishes no parameter list.
When we formed this view
Dates behind this page
Prices last checked 14h ago
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 5 listings publish no parameter list, so what their API accepts is unknown to us.
- Nothing we hold says whether an endpoint streams, so we do not show it either way.
- 2 of 5 listings do not say whether they train on prompts.
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-V3.2-Exp
- Architecture
- Mixture of experts
- Modality record
- text->text
- Catalogue slug
- deepseek-deepseek-v3-2-exp