DeepSeek-V3.2

DeepSeekOrganizationstablePopularity: 2796316
transformerssafetensorsdeepseek_v32text-generationbase_model:deepseek-ai/DeepSeek-V3.2-Exp-Basebase_model:finetune:deepseek-ai/DeepSeek-V3.2-Exp-Baselicense:miteval-resultsendpoints_compatiblefp8deploy:sagemakerregion:us
Context
131,072 tokens
Max Output
8,192 tokens
Released
12/1/2025
Providers
1

Overview

DeepSeek-V3.2 is a large language model developed by DeepSeek. It supports 131,072 tokens context and is available through 1 API provider(s).

Key Highlights

  • Unknown / Not disclosed

Brand

DeepSeek

Organization

Country: China

View all models, pricing and details from DeepSeek.

Specifications

Basic Information

ModelDeepSeek-V3.2
BrandDeepSeek
Version3.2
Model Typetransformers
Statusstable
Released12/1/2025
Popularity2796316

Architecture

Parameter Size685.4B
Architecture Typedeepseek_v32
Attention MechanismUnknown / Not disclosed
TokenizerUnknown / Not disclosed
Vocabulary SizeUnknown / Not disclosed

Context & I/O

Context131,072 tokens
Max Output8,192 tokens
ModalitiesUnknown / Not disclosed
LanguagesUnknown / Not disclosed

Capabilities

VisionNo
AudioNo
ReasoningNo
Function CallNo
MCPNo
JSON ModeNo
EmbeddingNo

Training Data

Training DatasetUnknown / Not disclosed
Estimated Data SizeUnknown / Not disclosed
Knowledge CutoffUnknown / Not disclosed

Inference & Deployment

API AccessYes
VRAM EstimateUnknown / Not disclosed
Quantization SupportUnknown / Not disclosed
Hardware RequirementsUnknown / Not disclosed

Safety & Compliance

Safety EvaluationsUnknown / Not disclosed
Content PolicyUnknown / Not disclosed
Region RestrictionsUnknown / Not disclosed

License & Pricing

Licensemit
Commercial UseUnknown / Not disclosed
Open SourceYes

Provider & Pricing

ProviderInputOutputCacheBatchWebsite
deepseek-ai$0.000/1M$0.000/1M--Visit Website

Price Changes

No recorded price changes.

Benchmarks

No benchmark data available.

Changelog

  • Page created / data imported
    9/23/2026
  • Last updated
    9/23/2026

Frequently Asked Questions

What is DeepSeek-V3.2?

DeepSeek-V3.2 is a large language model developed by DeepSeek.

Who developed DeepSeek-V3.2?

DeepSeek-V3.2 was developed by DeepSeek. You can learn more on the brand page or official website.

What is the context length of DeepSeek-V3.2?

DeepSeek-V3.2 supports up to 131,072 tokens of context.

How much does DeepSeek-V3.2 API cost?

API pricing for DeepSeek-V3.2 starts at $0.000/1M per 1M input tokens and $0.000/1M per 1M output tokens. See the pricing table above for provider-specific rates.

Related News

References

More from DeepSeek