Nvidia's Nemotron 3.5 Parse tops ParseBench for open-weight document parsing

AINvidia's Nemotron 3.5 Parse ranks first on ParseBench, a real-world document understanding benchmark covering more than 150 frontier, open-weight, and specialized models, according to Jerry Liu. The model has under 1B parameters, which the post says makes it very fast.

Jerry Liu

@jerryjliu0

View on X

Congrats to @nvidia for creating the best open-weight model on document parsing so far.

Nemotron 3.5 Parse takes the #1 spot on ParseBench, our comprehensive real-world benchmark for document understanding. The full list consists of 150+ frontier models, open-weight models, and specialized solutions.

It does this while remaining sub-1B parameters so it's also blazing fast ⚡️

Big thanks to @JensenHuang for the shoutout at GTC Berlin!

Nemotron release: https://huggingface.co/blog/nvidia/nemotron-retriever-world-class-agentic-retrieval

ParseBench: https://www.parsebench.ai/

Source details

Jerry Liu · x.comPublished Oct 21, 4:45 PM ET · added here 5:03 PM ET

Original title: Congrats to @nvidia for creating the best open-weight model on document parsing so far.

Nvidia's Nemotron 3.5 Parse tops ParseBench for open-weight document parsing

Topics