Skip to content
RankShark

DeepSeek V3.2 Exp

DeepSeek-V3.2-Exp is an experimental large language model released by DeepSeek as an intermediate step between V3.1 and future architectures

ModelOpen sourceby DeepSeeksince 2025
28.0

SharkScore

#340 in Models

Why it scores 28.0

DeepSeek V3.2 Exp has a SharkScore of 28.0/100, ranked #340 of 608 in Models (updated 2026-11-19).

Each part compares DeepSeek V3.2 Exp with other listings in Models. Evidence covers 87% of the score's weight.

Adoption
12 · 30%
User reviews
no data · 25%
Maintenance
2 · 20%
Trust & transparency
80 · 15%
Momentum
no data · 10%

Paying never changes this score. How ranking works · method v2026-11-18.2

RankShark review of DeepSeek V3.2 Exp

Automated evaluation by RankShark · 18 Nov 2026 · method v1.0 · How we review

We were not paid for this review.

As of 2026-11-18, DeepSeek V3.2 Exp is an experimental text model with a 163,840-token context window and measured OpenRouter rates of $0.27 in and $0.41 out per million tokens. [E1, E2, MEASURED]

DeepSeek-V3.2-Exp is an experimental large language model from DeepSeek. The vendor describes it as an intermediate step toward a future architecture and says it introduces DeepSeek Sparse Attention for long-context efficiency. [E1, E2]

Best for

  • People evaluating sparse-attention approaches for long-context text processing. · model card
  • Developers seeking local inference options through the listed Hugging Face demo, SGLang, or vLLM paths. · model card

Not for

  • Users who need image, audio, or other non-text input or output; the OpenRouter listing specifies text input and text output. · OpenRouter

Key features

  • DeepSeek Sparse Attention. The vendor says the model introduces fine-grained sparse attention to explore and validate training and inference efficiency improvements in long-context scenarios. · model card
  • Text input and output. The OpenRouter listing identifies the model's input and output modalities as text. · OpenRouter
  • Context window. RankShark measured a 163,840-token context window on the OpenRouter listing on 2026-11-18. · OpenRouter
  • Local inference options. The vendor provides an inference demo and describes launch paths for SGLang and vLLM. · model card
  • Model weights license. The vendor says the repository and model weights are licensed under MIT. · model card

Pros

  • The vendor provides local inference instructions for Hugging Face, SGLang, and vLLM. · model card
  • The vendor says the repository and model weights use the MIT License. · model card
  • The OpenRouter listing states a 163,840-token context window; RankShark measured the same listing value. · OpenRouter

Cons

  • The vendor labels this an experimental release and an intermediate step toward a future architecture. · model card
  • The OpenRouter listing specifies text-only input and output modalities. · OpenRouter
  • No free plan or free trial stated on the pages we checked.
  • Whether customer data is used for model training is not stated on the pages we checked.

Pricing

PlanPriceLimits
DeepSeek: DeepSeek V3.2 Exp on OpenRouter$0.27 per million input tokens; $0.41 per million output tokens163,840-token context window

Checked 18 Nov 2026

Free plan
Not stated (checked 18 Nov 2026)
Free trial
Not stated (checked 18 Nov 2026)
Trains on customer data
Not stated (checked 18 Nov 2026)
Data retention
Not stated (checked 18 Nov 2026)
Data regions
Not stated (checked 18 Nov 2026)
Data processing agreement
Not stated (checked 18 Nov 2026)
AI models used
DeepSeek-V3.2-Exp · model card, OpenRouter
API
OpenRouter lists chat completions, completions, messages, and responses endpoints. · OpenRouter
SDKs
Not stated (checked 18 Nov 2026)
Webhooks
Not stated (checked 18 Nov 2026)
Platforms
The vendor documents local inference options for Hugging Face, SGLang, and vLLM. · model card
Security claims
No security certification stated (checked 18 Nov 2026)

What we measured

Context window (OpenRouter listing) · now
163,840 tokens
Price per million tokens (OpenRouter listing) · now
$0.27 in, $0.41 out
Package downloads · 30 days
54,829
How we tested · sources
  • Hourly uptime and response-time checks of the product's URL.
  • Repository, release and download data from GitHub, npm, PyPI and Hugging Face where they exist.
  • We did not sign up, pay, or use the product as a customer.
  • 18 Nov 2026: Plans or prices changed on the pricing page.

Something wrong? Report an error with a public link that shows it; corrections re-run the review.

Top alternatives to DeepSeek V3.2 Exp

See all
  • 67.1

    Qwen3.8 27B is an open-weight dense vision-language model from Qwen.

    ModelOpen source
  • 65.5

    Qwen3.6-35B-A3B is an open-weight multimodal model from Alibaba Cloud with 35 billion total parameters and 3 billion active parameters per t

    ModelOpen source
  • 64.7

    Qwen3.8 Flash is a multimodal reasoning model from Alibaba.

    ModelOpen source
  • 64.3

    Qwen3.6 27B is a dense 27-billion-parameter language model from the Qwen Team at Alibaba, released in April 2026.

    ModelOpen source