DeepSeek-V3.1 is a large language model released by DeepSeek, released under the mit license, with a claimed knowledge cutoff of 2025-03-31. Its weights, configuration, and tokenizer files have been cryptographically hashed and are queued for on-chain attestation by Weight Registry.
Description
DeepSeek-V3.1 is a large hybrid reasoning model (671B parameters, 37B active) released August 2025, supporting both thinking and non-thinking modes via prompt templates. It extends the DeepSeek-V3 base with a two-phase long-context training extension.
Metadata
Hugging Face stats
319,041 downloads on Hugging Face · 833 likes · pipeline: text-generation · library: transformers
Weight Hashes
Weights file
SHA-2564262ef9f2a11a19d14336787c3d59f2203fcddec9ac0d929e8e853cc4b3c61e7
SHA3-256839a41233f9dc58381e3252b996072d83d314941174dbfae390ae20bc34539e3
BLAKE38d0aa5e6af8bdd66541264a8db7980dab11690e7b25e3bb158df439a71122ade
Config file
SHA-2563e5d192d745758862b0a3e6338ee181991db49933dbbd3185e1bde34944e7bad
SHA3-25666e190eb43cbcae58d92d73171eb2e2dcf416da8256795a09ed86545e7cdf6b9
BLAKE308022107b9235a1cc4cfec70fc5996f8982adb2abd3ecb2db89fda9509d6aff5
Tokenizer file
SHA-25632b34a41212e92f62e859cbbea121ae705a1fabbf157d9acf22d134ecd8dcf70
SHA3-25695c4e528ed7d934348da7b77e5b148f03376fe60ae8f6b42db0edfb89a27621d
BLAKE38a494a27103889d2ba0707bf0437df5b4707c55666377c0267cf3ece05fe352c
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.