15.9M-parameter OCR model Kraken PP-OCRv6 beats VLMs hundreds of times larger
vanstriendaniel · x · 2026-09-03
On a 2,165-page historical-document eval, the 15.9M-parameter specialist OCR model Kraken PP-OCRv6 ranks #4 on reading CER and #1 when long s, ligatures and case are preserved, beating general-purpose VLMs hundreds of times its size. The author notes the model did so well the plotting logic had to be adjusted.
Related event: Tiny 15.9M-parameter Kraken OCR model beats far larger VLMs(2 posts)→
More from Models
- China's token plans priced far above overseas LLM subscriptions, per finance mag piece — yihui_indie · 2026-09-03
- Hands-On Ranking: Fable 5.1 Beats Claude Opus on RL Environment Creation — HarveenChadha · 2026-09-03
- Philosopher says Fable 5.1 found Arrhenius's favored population ethics impossibility theorem is false — willmacaskill · 2026-09-03
- Reddit users report mass OpenAI account bans hitting paying customers — Existing-Slide7395 · 2026-09-03
- Gemini Speech-to-Text Shines in Real-World Test—Dev Open-Sources Windows Tool AivoRelay — lvvy · 2026-09-03
- Meta's new Muse model suddenly outputs Chinese characters, sparking distillation rumors — zsakib_ · 2026-09-03