AI article
I Benchmarked My AI Carb Counter on 400 Meal Photos: 8 Models, 1 Holdout Set
Community description: I build Soba, an iPhone app that looks at a photo of a meal and estimates the carbs in it. People...
Dev.to | Oct 8, 2026 | Maksim Danilchenko
Automated excerpt
The main metric is the share of meals within 10 g of the reference carbs, roughly one bread unit. Letting Gemini 3. 6 Flash reason longer raised the share within 10 g by 4. 1 points over all 400 meals, but the share within 5 g and the number of errors above 30 g stayed the same, and each scan took 6. 6 seconds instead of 3. 5. Even depth left 12 of 13 large meals underestimated.
Selected automatically from source text; not independently written or fact-checked. Read the original for full context.