Model
DeepSeek V3.2
Recorded as suitable for chat, rag, coding, reasoning, multilingual, zh.
- Source
- Model registry API
- Verified
- 19 Sep 2026
- Confidence
- High
01What this is
- Author
- deepseek-ai
- Family
- DeepSeek V3
- Task
- text-generation
- Hugging Face
- deepseek-ai/DeepSeek-V3.2
- OpenRouter
- deepseek/deepseek-v3.2
- Also known as
- DeepSeek V3.2, deepseek-v3.2, deepseek v3
02Registry facts
- LicenceFACT
mit
- FreshRetrieved 19 Sep 2026
- ParametersFACT
685.4B
- FreshRetrieved 19 Sep 2026
- Context lengthFACT
not returned by the registry
- Evidence not verified
- Downloads (30 days)FACT
2,421,883
- FreshRetrieved 19 Sep 2026
- LikesFACT
1,489
- FreshRetrieved 19 Sep 2026
- GatedFACT
no
- FreshRetrieved 19 Sep 2026
Model
- downloadsFACT
2421883
- Verification may be outdatedRetrieved 19 Sep 2026
- gatedFACT
false
- Verification may be outdatedRetrieved 19 Sep 2026
- hosted context length · SiliconFlowFACT
163840
- Verification may be outdatedRetrieved 19 Sep 2026
- hosted context length · StreamLakeFACT
128000
- Verification may be outdatedRetrieved 19 Sep 2026
- hosted context length · GMICloudFACT
163840
- Verification may be outdatedRetrieved 19 Sep 2026
- likesFACT
1489
- Verification may be outdatedRetrieved 19 Sep 2026
- model licenseFACT
mit
- Verification may be outdatedRetrieved 19 Sep 2026
- parameter countFACT
685355329792
- Verification may be outdatedRetrieved 19 Sep 2026
Pricing
- hosted price completion per m · SiliconFlowFACT
0.4200
- AgeingRetrieved 19 Sep 2026
- hosted price completion per m · StreamLakeFACT
0.3217
- AgeingRetrieved 19 Sep 2026
- hosted price completion per m · GMICloudFACT
0.3096
- AgeingRetrieved 19 Sep 2026
- hosted price prompt per m · SiliconFlowFACT
0.2590
- AgeingRetrieved 19 Sep 2026
- hosted price prompt per m · StreamLakeFACT
0.2145
- AgeingRetrieved 19 Sep 2026
- hosted price prompt per m · GMICloudFACT
0.2088
- AgeingRetrieved 19 Sep 2026
Other
- available on platform · azureFACT
azure
Evidence not verified — the quoted phrase was not found in the fetched document.
- Evidence not verified
- available on platform · awsFACT
aws
The source document changed after this was recorded; it is queued for re-checking.
- Verification may be outdatedRetrieved 19 Sep 2026
- last modifiedFACT
2025-12-01T11:04:59.000Z
- Verification may be outdatedRetrieved 19 Sep 2026
- libraryFACT
transformers
- AgeingRetrieved 19 Sep 2026
03VRAM by quantisation
Peak VRAM for serving this model, computed by lib/deployment/hardware.ts from weights + KV cache + 15% runtime overhead. Inputs: 685.4B parameters, 8,192 tokens of context (assumed — none fetched yet), 4 concurrent requests (assumed).
| Quantisation | Weights | Total |
|---|---|---|
| bf16bf16 is two bytes per weight, the same arithmetic as fp16. | 1370.71 GB | 1593.6 GB |
| fp8 | 685.36 GB | 805.45 GB |
| int4 | 342.68 GB | 411.37 GB |
What this estimate assumes
- Weights: 685.4B parameters × 2 bytes per parameter (fp16) = 1370.71 GB.
- No model configuration was supplied, so the KV cache uses the size heuristic for this band: 0.459 MB per token, from extrapolated above 90B — no configuration was fetched for this band.
- KV cache: 0.459 MB per token × 8,192 tokens of context × 4 concurrent requests = 15.03 GB, sized for every user holding a full context at once.
- Overhead: 15% of weights plus KV cache for the runtime, activations and memory fragmentation = 207.86 GB.
- GB means 10⁹ bytes, matching how GPU memory is advertised.
- KV cache held at fp16 (2 bytes per element); quantising weights does not by itself quantise the cache.
- Peak figure, not average: prefix caching and paged attention usually keep real usage lower, while speculative decoding, CUDA graphs and a second resident model push it higher.
- "Concurrent" counts requests generated at the same instant, not people: 4 simultaneous requests is the worst case this estimate is built on — measure yours before buying.
04Hardware profiles
No hardware profile was sized for this answer.
05Hosted providers
- Prompt (per million tokens, USD)FACT
0.20879999999999999
Source: openrouter
- FreshRetrieved 19 Sep 2026
- Completion (per million tokens, USD)FACT
0.30960000000000004
- FreshRetrieved 19 Sep 2026
| Provider | Context |
|---|---|
| GMICloud | 163,840 |
| StreamLake | 128,000 |
| SiliconFlow | 163,840 |
| DeepInfra | 163,840 |
| AtlasCloud | 163,840 |
| Venice | 160,000 |
06Deployment stacks
No published deployment stack names this model yet.
07Evidence
- 01Tier 2official vendor legal documentationVendor legal and technical documentation
Models at a glance - Amazon Bedrock
https://docs.aws.amazon.com/bedrock/latest/userguide/model-cards.html
AgeingRetrieved 19 Sep 2026sha256:4c3b2b5762261 record
- 02–10Tier 4official repository / model cardOpenRouter
OpenRouter endpoints — deepseek/deepseek-v3.2
https://openrouter.ai/api/v1/models/deepseek/deepseek-v3.2/endpoints
Verification may be outdatedRetrieved 19 Sep 2026sha256:1c88b0d031879 records · · · · · · · ·
- 11–17Tier 4official repository / model cardHugging Face Hub
Hugging Face model card data — deepseek-ai/DeepSeek-V3.2
https://huggingface.co/api/models/deepseek-ai/DeepSeek-V3.2
Verification may be outdatedRetrieved 19 Sep 2026sha256:af13f68f2fef7 records · · · · · ·
Improve this page
Sign in to contribute
From the field
0 deployments · 0 questions
Nobody has reported deploying this here yet, and no question has been opened against this page. Both appear once a reviewer accepts them.