NVIDIA's sub-1B open-weight Nemotron 3.5 Parse takes #1 on LlamaIndex's ParseBench
lmoroney · x · 2026-10-23
- Per Jerry Liu, NVIDIA's Nemotron 3.5 Parse claimed the top spot on ParseBench, LlamaIndex's real-world document understanding benchmark covering 150+ frontier, open-weight and specialized models.
- The model is open-weight and under 1B parameters, making it fast; it shipped alongside new Nemotron Embed and Rerank models.
- The author notes parsing is where much RAG quality quietly gets lost, and suggests benchmarking it against your current pipeline on messy, table- and chart-heavy PDFs.
More from Infra
- Meta researcher: horrifying infra, not talent, is why OSS can't catch closed frontier — A_K_Nain · 2026-10-23
- Jane Street and NVIDIA speed up Mamba-2 with custom cuTile kernels — blelbach · 2026-10-23
- LatentLoop: asymmetric compute allocation cuts looped-MoE training cost by 19% — pmttyji · 2026-10-23
- Same AI request emits 8.5x more CO2 in Texas than in Sweden — jlundberg · 2026-10-23
- SK Hynix Pulls Plug on Solidigm US IPO as Chip Stocks Slide, Shares Down ~30% This Quarter — rohanpaul_ai · 2026-10-23
- Designer mocks Nvidia: buy idle GPUs or lose access to new ones—definitely not a bubble — michalmalewicz · 2026-10-23