Qwen3-4B-Instruct-2507 is a large language model released by Alibaba, released under the apache-2.0 license, with a claimed knowledge cutoff of 2025-07-01. Its weights, configuration, and tokenizer files have been cryptographically hashed and permanently attested on Ethereum (Sepolia) by Weight Registry.
Description
Dated mid-generation refresh of the Qwen3 4B instruct checkpoint.
Metadata
Hugging Face stats
3,964,449 downloads on Hugging Face · 976 likes · pipeline: text-generation · library: transformers
Weight Hashes
Combined commitment: a Merkle root over the weights, config, and tokenizer hashes together (not a raw weights-only hash) -- see this project's v2 on-chain schema.
Combined SHA-2560x4fe09f97fe0b4ff8ee014f8620bc28f30c53501e2ced433b0763207626b1aa3f
Combined SHA3-2560x9a67b839e754cef6627857db77c54aa56baed5f50297acc6853130a1a4ffca52
Combined BLAKE30x287cc4d005bb1b773a0b31a1caa0fa893eee88610e85619f163d377b405acea1
Per-file breakdown
Computed locally prior to attestation; the combined hash above is the canonical on-chain commitment.
Config file
SHA-2565beea1a4a34c62782bfb2f911c606741a3bab8f92d80a118fa053c28af12e8ba
SHA3-25674bbcd42e65a64af36fc373b94edbdb99f806078282f84f0a79a610bd428c532
BLAKE37ee8c9dbd8a9f02c11b9fd0267f566fb90c46696a15faca9b9a8f2fde7ec2671
Tokenizer file
SHA-256aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4
SHA3-256427a6d44f980384d37f957696c4f7d191d0e3b3b7819bf4ee5c84634dd86a4a0
BLAKE3b0cb923fc505fdf0a53f0287654fa26577d3f333d4134350da0a97664b228739
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.