With --math flag, 12M-param Indra OCR scores 71.4 on OmniParseBench, beating HunyuanOCR's 64.6
VikParuchuri · x · 2026-10-20
Author VikParuchuri shares eval details: with the --math flag, Indra OCR scores 71.4 on OmniParseBench, beating the 1.1B-param HunyuanOCR (64.6) and closing in on the 3.4B-param DeepSeek OCR 2 (75.2). The flag routes only math content to a larger-parameter model, bringing total params to 12M. A scatter chart shows Indra as the smallest model on the open-weight OmniParseBench frontier.
More from Models
- Report: Liang Wenfeng planned independent DeepSeek compute chips before Huawei persuaded him otherwise — teortaxesTex · 2026-10-20
- Smallest AI's full voice stack hits 26.8% completion at $0.048/call on 300 hard calls — auto_grad_ · 2026-10-20
- NVIDIA's Dory1.0-4B Looped Model Lets You Pick Inference Depth at Runtime — lmoroney · 2026-10-20
- OpenAI dropped 722 AI-generated math papers in a day; can humans read them? — songyoupeng · 2026-10-20
- JevBench adds 34 decision models in four days; Nimble-v4-max price cut to $0.04/1M tokens — airesearch12 · 2026-10-20
- Mercury 3 benchmarks: beats GPT-6 Luna and Haiku 5.5 at 1,068 tok/s — StefanoErmon · 2026-10-20