EgoPoint-Bench - Simulation OCR: leaderboard

EgoPoint-Bench (2026) asks multimodal models about the object a first-person hand points at, on held-out simulation scenes and real smart-glasses recordings, zero-shot; true/false and single-choice answers by exact match and open answers scored by a GPT-4o judge. This board covers the OCR questions (text on the target) among the 1,929 simulation test questions. Accuracy in percent; higher is better.

Metric: Accuracy (%) on the OCR questions among the 1,929 held-out simulation test questions (HM3D, HSSD and AI2-THOR scenes); zero-shot (direct inference); true/false and single-choice questions by exact match, open-ended answers scored 0 to 1 by a GPT-4o judge; greedy decoding for open models, temperature 0 for API models; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 12 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Qwen 3 VL 32B Instruct79.39#606
2Qwen 3 VL 8B Instruct77.58#1029
3GPT-5.2 Instant75.45#304
4Gemini 3 Pro74.85#145
5Gemini 3 Flash73.64#175
6InternVL3.5-14B-Instruct71.52#1736
7GPT-5 Mini68.79#364
8InternVL3.5-2B-Instruct66.97#6517
9InternVL3.5-8B-Instruct66.36#1891
10GLM-4.6V-Flash61.52#1040
11LLaVA-NeXT-7B53.33#5032
12LLaVA-v1.5-7B50.91#7630

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=egopoint-bench-simulation-ocr · How It Works · Data refreshed daily, snapshot 2026-11-21.