Evidence checked Nov 9, 2026

Verified Nov 9, 2026 via DeepSeek V3: deepseek-v3-projectAI models

DeepSeek V3

DeepSeek describes V3 as a mixture-of-experts language model with 671 billion total and 37 billion activated parameters.

Verified facts
8
Offers
1
Sources
2
Pricing
Free offer

Overview

Model scope

DeepSeek describes V3 as a mixture-of-experts language model with 671 billion total parameters and 37 billion activated for each token. The project lists a base model and a chat model; this profile treats the family as one product, not separate entries for variants.

The project documentation describes model downloads and deployment paths. Confirm the exact model identifier, current access, hardware requirements and service terms for the way you intend to use it. The supplier’s parameter figures are not an independent performance result.

Read the full editorial overview

Licences and evidence

The captured repository distinguishes an MIT licence for code from a separate model licence for the weights. Review the exact model agreement, dependencies, hosted-service terms and trademarks rather than assuming the code licence covers every use.

No model inference, API request, account, benchmark, hardware deployment or performance test was performed. The selected material does not establish a comparable current hosted token price or service guarantee.

Sources

DeepSeek V3: deepseek-source · Evidence: documentation · Not checked

Plans & prices

Prices as listed by DeepSeek V3, before any tax or fee the pricing page does not state.

Scroll the table sideways for every column.

PlanRate period not statedCoverage limits
Plan: Self-hosted
FreeTax: Confirm at checkoutUpfront: not established
ConditionsDeepSeek License Agreement 1.0 (LICENSE-MODEL at commit 9b4e978, captured 9 November 2026) grants a perpetual, worldwide, no-charge, royalty-free copyright license to the model weights; the README says the code is MIT-licensed and that DeepSeek-V3 Base/Chat supports commercial use. Use-based restrictions (Attachment A) and redistribution conditions apply. Applies to the downloaded open weights run on your own infrastructure; hardware, hosting and third-party services are separate. DeepSeek's chat website and API (pricing not captured here) are separate offerings. Supplier licence text, not legal advice.
Unknown fees: Infrastructure and third-party servicesVerified Nov 20, 2026
Local deployment is listed with the DeepSeek-Infer demo, SGLang, LMDeploy, TensorRT-LLM, vLLM, LightLLM, AMD GPUs and Huawei Ascend NPUs, in FP8 and BF16 modes (TensorRT-LLM: BF16 and INT4/8; Ascend: INT8 and BF16).
ConditionsPublisher repository README at commit 9b4e978 captured 9 November 2026; supplier description, not a deployment, inference or independent test. Only FP8 weights are published and a conversion script produces BF16; the demo supports Linux with Python 3.10 only (Mac and Windows are not supported) and Hugging Face Transformers is not directly supported. Some framework features are marked coming soon. GPU counts and memory needs are not stated in the README.
Verified Nov 9, 2026

DeepSeek V3: deepseek-v3-modellicense-project · DeepSeek V3: deepseek-v3-project

Documented facts

Each fact comes from official documentation and shows when it was verified.

  • Who it is for

    DeepSeek describes V3 as a mixture-of-experts language model with 671 billion total and 37 billion activated parameters.

    Verified Nov 4, 2026 · DeepSeek V3: deepseek-source

  • Coverage limits

    The project states that its code and model weights are governed by separate licences.

    Verified Nov 4, 2026 · DeepSeek V3: deepseek-source

  • Provider

    DeepSeek

    Provider named by the DeepSeek License Agreement 1.0 definition of "DeepSeek" (LICENSE-MODEL at commit 9b4e978, captured 9 November 2026); the legal entities are Beijing and Hangzhou DeepSeek Artificial Intelligence Fundamental Technology Research Co., Ltd. and affiliates.

    Verified Nov 9, 2026 · DeepSeek V3: deepseek-v3-modellicense-project

  • API model identifier

    DeepSeek-V3

    Verified Nov 4, 2026 · DeepSeek V3: deepseek-source

  • Context window

    128,000 tokens in the context window

    Publisher repository README at commit 9b4e978 captured 9 November 2026; supplier description, not a deployment, inference or independent test. Context length listed for DeepSeek-V3 and DeepSeek-V3-Base in the Model Downloads table; the README's needle-in-a-haystack tests cover context lengths up to 128K. The hosted service may differ.

    Verified Nov 9, 2026 · DeepSeek V3: deepseek-v3-project

  • Availability

    Available on DeepSeek's official chat website and through an OpenAI-compatible API on the DeepSeek Platform.

    Publisher repository README at commit 9b4e978 captured 9 November 2026; supplier description, not a deployment, inference or independent test. The README does not say which model version those hosted services currently run, and their prices and limits are not captured here; open weights are covered by the separate Self-hosted plan facts. Account, region and platform eligibility require confirmation.

    Verified Nov 9, 2026 · DeepSeek V3: deepseek-v3-project

Applies to the facts above: Publisher repository README at the captured revision, 4 November 2026. A named variant is evidence within the model family, not a separate product. No model inference, API call, account, benchmark or performance test. Current availability and terms require separate review.

Not confirmed in our sources

We have not confirmed these terms for DeepSeek V3 in the linked documentation. That does not mean DeepSeek V3 lacks them; confirm them before you rely on them.

  • Modalities
  • Maximum output
  • Input token rate
  • Output token rate

Sources & methodology

Facts come from the linked documentation. Documentation checks are not hands-on tests, and readable sources can have incomplete coverage.

DeepSeek V3: deepseek-v3-projectraw.githubusercontent.com
Retrieved Nov 9, 2026
Retrieved Nov 9, 2026

Discovered an undocumented discrepancy or pricing shift? Submit a correction request.

Submit correction

DeepSeek V3 alternatives

Other AI models tools to weigh. Different billing units do not establish a cost winner, so compare plans side by side.

DeepSeek V3 comparisons

Compare two products in the same category using their documented terms and sources.