← Back to Explorer
Pending

DeepSeek-V4-Flash (2026)

DeepSeek-V4-Flash is a large language model released by DeepSeek, released under the MIT license, with a claimed knowledge cutoff of 2026-01-01. Its weights, configuration, and tokenizer files have been cryptographically hashed and are queued for on-chain attestation by Weight Registry.

Description

Faster/smaller variant in DeepSeek's V4 generation, part of the lab's continuing dense release cadence.

Metadata

Type
LLM
Issuer
DeepSeek
License
MIT
Size
148.66 GB
Knowledge cutoff
2026-01-01
Source
https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash
Revision
60d8d70770c6776ff598c94bb586a859a38244f1

Hugging Face stats

1,307,475 downloads on Hugging Face · 2,259 likes · pipeline: text-generation · library: transformers

transformerssafetensorsdeepseek_v4text-generationeval-resultsendpoints_compatible8-bitfp8

Weight Hashes

Weights file
46 shards (Merkle root) · 148.66 GB
SHA-256e28998c9e6f23a20d37004392d106208cb4bbbc370e24ee2fccce1adf816ba4f
SHA3-256c461012eb809afef42ad2fd5cfe3bc7102bb05f37e6eb9f871e598062556ea9e
BLAKE3a610f38ae059078ca7739f5ea85f5796381165f45da0c02ec0720bffedae7140
Config file
config.json · 1.71 KB
SHA-256b628e63398a645abc711d92207f8737dd8140f7a4ef1e0a5b3616019e0ddd818
SHA3-256a37b822e7fb83c7e3e1226ab613be81b136b6726ac1f882330f66e62fcb8c449
BLAKE3ac4b0c2c93ce417e87dfebfe800e82112ded6a322cce28fd10dbbf3873f9989e
Tokenizer file
tokenizer.json · 6.07 MB
SHA-2568f9f37ca37fdc4f5fd36d5cf4d3b0e8392edb4e894fd10cc0d70b4957c8633cf
SHA3-256ca52f8556ff84e23db6fb269a9afe9b5915db355725a809ee5258aef29caf304
BLAKE37ebd3b6f0d2fca9a770b2a8c190b4d815684735219c294d7e5dcf314ed6a4eb4

Verify it yourself

Download the file(s) from the source below, then hash them locally (sha256sum <file> on Linux/macOS, certutil -hashfile <file> SHA256 on Windows) and compare the output against the hash(es) shown above.