# deepseek-ai/DeepSeek-V2

DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model

Repository: https://github.com/deepseek-ai/DeepSeek-V2
Canonical: https://ross.abutalabs.com/products/deepseek-v2
License: MIT
License Family: permissive
Last push: 2024-09-25T10:23:55+00:00

## Health v2 (maintenance only)
Score: 24/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 61
- inputs: {"age_days": 863, "days_push": 707, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 5039, forks 550 (observed 2026-08-28T04:09:08.767062+00:00)

## What it is
DeepSeek-V2 is a strong, economical Mixture-of-Experts language model released with open weights, inference code, and evaluation tooling. The repository provides model downloads, architecture details, and API access for the 236B-parameter MoE model family.

## Use cases
- run a large mixture-of-experts language model locally
- download open weights for a strong open-source LLM
- serve a cost-efficient MoE chat model via API
- evaluate DeepSeek-V2 on benchmarks
- fine-tune or build on an open MoE base model

## When to choose
- you need an open-weights, high-capability MoE LLM with efficient inference
- you want to self-host a strong chat model on GPU infrastructure
- you're researching mixture-of-experts architectures

## When to avoid
- you need a small model that runs on consumer hardware without large GPU memory
- you only want a hosted API without managing model weights
- you need a permissively licensed model for commercial redistribution without checking the model license

## Facets
- artifact type: library
- maturity: active
- function: llm-inference, machine-learning, deep-learning
- domain: large-language-models, deep-learning, artificial-intelligence
- platform: python
- tags: mixture-of-experts, open-weights, transformer, huggingface, model-weights, gpu, linux

## Member repositories
- deepseek-ai/DeepSeek-V2 (main) score 24

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:08.767062+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:17:24.753178+00:00, confidence not recorded.
  - readme: https://github.com/deepseek-ai/DeepSeek-V2 (fetched 2026-08-28T04:09:08.767062+00:00, sha 9224fda4c8d3)
- Data as of 2026-08-30T08:39:29.467469+00:00.