Qwen3.8 27B is an open-weight dense vision-language model from Qwen.
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
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
| Plan | Price | Limits |
|---|---|---|
| DeepSeek: DeepSeek V3.2 Exp on OpenRouter | $0.27 per million input tokens; $0.41 per million output tokens | 163,840-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.
- model card: https://huggingface.co/deepseek-ai/DeepSeek-V3.2-Exp · fetched 18 Nov 2026
- OpenRouter: https://openrouter.ai/deepseek/deepseek-v3.2-exp · fetched 18 Nov 2026
- 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- 67.1ModelOpen 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