16 Open Models Read a Fax-Blurred Dosing Table: 7 Misread 300mg as 100mg, None Flag It
MaziyarPanahi · x · 2026-11-06
Maziyar Panahi tested open models on children's dosing tables using the RxHandBD handwritten prescription dataset (5,578 word images, CC BY 4.0) and FDA labels from DailyMed.
On a clean paper dosing table, all 16 top open models correctly read 300 mg. After a simulated fax degradation, the median model's accuracy fell from 93% to 85%, and 7 models misread the broken digit as 100 mg — none of the 7 flagged uncertainty.
His takeaway: "a wrong number in a neat table is worse than a blurry page" — in medical dosing, fluently outputted errors are more dangerous than admitted illegibility. Best first-run model was Muse Spark at 78.3; topiramate 12–22 kg rows were the riskiest after faxing. Not medical advice.
More from Models
- From "can't write coherent text" to complaining a 2B model can't run a training pipeline in 10 minutes — cephaloform · 2026-11-07
- Rubricon: scaling test-time compute to auto-generate expert-grade LLM rubrics — deliprao · 2026-11-07
- Kimi K3.1 Pro lands on Microsoft Foundry with 1M-token context for agentic work — DanWahlin · 2026-11-07
- Ripple 65B Preview vs Qwen 3.8 coding benchmarks: 79.1 on SWE-Bench Pro — Nota_ReAlperson · 2026-11-07
- Fireship: OpenAI's unnamed model just cracked 3-SAT, threatening a 27-year-old hypothesis — Fireship · 2026-11-07
- Bittensor subnet model tops respected independent leaderboard just ~30 days after launch — airesearch12 · 2026-11-07