SYNCR - Route Planning: leaderboard

Accuracy (%) on planning a route from fragmented egocentric videos in SYNCR (2026), a synthetic benchmark of multi-video questions with programmatically verified answers; zero-shot four-option multiple choice, exact match, 200 questions per task (chance 25). Higher is better.

Metric: Exact-match accuracy (%) on the SYNCR Route Planning task (planning a route from fragmented egocentric videos), zero-shot four-option multiple choice over 2 to 4 labelled input videos, exact match, deterministic decoding, on the paper's balanced subset of 200 questions per task; chance is 25; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 18 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1GPT-5.444#135
2Gemini 3 Flash36#175
3LLaVA-OV-72B33.5#1879
4InternVL3.5-14B30.5#1504
5Qwen 3 VL 32B Instruct29.5#606
6InternVL3.5-38B27.5#1021
7Qwen 2.5 VL 72B Instruct22.5#841
8Qwen 2.5 VL 7B Instruct21.5#1902
9Qwen 2.5 VL 32B Instruct21.5#1062
10LLaVA-OV-7B21.5#4475
11InternVL3.5-4B17.5#1930
12Qwen 3 VL 2B Instruct17.5#2367
13InternVL3.5-8B16#1604
14LLaVA-OV-0.5B15#4352
15Qwen 3 VL 8B Instruct14.5#1029

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

Interactive version: theaggregate.ai/benchmark?slug=syncr-route-planning · How It Works · Data refreshed daily, snapshot 2026-11-21.