Egocentric annotation cost scaling: Perceptron Ego 1.0 sits on the Pareto frontier
AkshatS07 · x · 2026-11-11
Using EgoHome, the author measured quality-vs-cost scaling for egocentric subtask annotation. Perceptron Ego 1.0 at 4 FPS ($9/video-hour, score 0.787) and 8 FPS ($15, 0.798) sits on the Pareto frontier, outpricing rivals on quality: GPT-6 Astra 4 FPS ($76, 0.797), GPT-6 Astra 1 FPS ($20, 0.775), NVIDIA FLORA ($35, 0.747), Glimpse 4 FPS ($2.8, 0.689) and Gemini 3.8 Flash 4 FPS ($1.3, 0.663). Cheap-end picks remain Gemini 3.8 Flash and Glimpse.
Related event: Perceptron Releases Ego 1.0 Egocentric Annotation System(2 posts)→
More from Research
- MKA deep dive: top models cluster at 95-97% on basics but split 28-57% on hard tasks — ArtificialAnlys · 2026-11-11
- MKA Benchmark by Output Type: No Model Tops 52% on Word or Text Files — ArtificialAnlys · 2026-11-11
- MKA Vertical Results: No Single Model Dominates Across 10 Industries — ArtificialAnlys · 2026-11-11
- Google DeepMind Launches MKA Benchmark: One Task Cost Claude $68 Across 224 Turns — ArtificialAnlys · 2026-11-11
- DeepMind-backed MKA benchmark crowns Claude Fable 5.2 at 57%, with 30x cost spread — ArtificialAnlys · 2026-11-11
- Emergent misalignment is predictable OOD; later-layer fine-tuning reduces transfer — AdtRaghunathan · 2026-11-11