`deepseek-ai/DeepSeek-V3.2`

mit 1 launch

text-generation · 685.36B parameters · 642.1 GiB of weights at revision a7e62ac0. Needs at least 8 GPUs.

download the card performance the weights on HF

Did this card launch? The count is real — it comes from launches the kit recorded against this path, not from a widget anywhere.

1checkpoints

1published launches

8GPU minimum

1named gaps

The identity, read from the HF API at the revision the card pins.
weight path	`deepseek-ai/DeepSeek-V3.2`
revision	`a7e62ac04ecb2c0a54d736dc46601c5606cf10a6`
license	mit
gated	no
parameters	685.36B
weight bytes	642.1 GiB
modality	text-generation
HF downloads	6.4M at capture
last modified	2025-12-01T11:04:59.000Z

The checkpoints.

Every servable variant in or beside the repo. resident_gib is what lands in VRAM after load; host_gib is what stays on the host and never touches the card.

name	format	location	resident	host tier	gated
`root`	bf16	repo root	642.1 GiB	—	no

root — 163 weight file(s)

The published launches.

One entry per engine × checkpoint with a command somebody published. Each cites the page it was read from — the card copies it, it does not invent it.

vllm `root`

vllm/vllm-openai:latest

argv — what a wrapper execs

vllm serve deepseek-ai/DeepSeek-V3.2 --trust-remote-code --disable-flashinfer-autotune --tensor-parallel-size 8 --tokenizer-mode deepseek_v32 --tool-call-parser deepseek_v32 --enable-auto-tool-choice --reasoning-parser deepseek_v3

as the source printed it

vllm serve deepseek-ai/DeepSeek-V3.2 \
  --trust-remote-code \
  --disable-flashinfer-autotune \
  --tensor-parallel-size 8 \
  --tokenizer-mode deepseek_v32 \
  --tool-call-parser deepseek_v32 \
  --enable-auto-tool-choice \
  --reasoning-parser deepseek_v3

validated on: h100, h200
env: VLLM_USE_RUST_FRONTEND=1
source: vllm-recipes

Every floor a box must clear.

A launch that fails on the third requirement wasted the first two. These are the requirements, as data — the fit gate reads them instead of a paragraph.

The hardware block, verbatim from the card.
GPU models	h100, h200
architectures	hopper
min GPUs	8
TP rule	widest published TP is 8
min VRAM / GPU	—
min total VRAM	—
min disk	821 GiB
min host RAM	—
interconnect	not required
CUDA min	—
driver min	—
context tokens	—

The 1 thing this card could not answer.

First-class, not swallowed. Every entry is a field a source did not answer — which makes it a launch nobody has done, and therefore a unit of work rather than a defect.

envs.sglang no SGLang cookbook page matched this repo

The file.

The card is a YAML document with a published JSON Schema. View it, download it, diff it against the next capture — and point one at it.

download

launch-card.yaml card.json the schema

Same bytes the collector wrote. The JSON is the same card, for a machine that would rather not parse YAML.

with one

one launch --card deepseek-ai/DeepSeek-V3.2/launch-card.yaml

one follows the argv above, checks the floors before the paid clock, and on a failure changes exactly one field and retries — recording every attempt against this path.

Where every field came from.

The card is assembled, not authored. These are the sources, so a reader can check the claim rather than trust it.

source	url
hf-api	`https://huggingface.co/api/models/deepseek-ai/DeepSeek-V3.2`
vllm-recipes	`https://recipes.vllm.ai/deepseek-ai/DeepSeek-V3.2.json`

Captured 2026-10-31T17:28:25Z by hf-api + vllm-recipes + sglang-cookbook. Schema onefrontier.launch-card/v1.