Tag · 9 charts · newest first
Computer Vision
Our charts and analysis on computer vision: how well AI models understand images, diagrams, clocks and calendars, and video, on tests such as MMMU, Video-MMMU and the Vision Arena. The charts are under AI Image Understanding.
-
RoboticsNvidia's 2025 PhysicalAI-AV dataset holds 1,727 hours of driving, the most of 13 benchmarks tracked19 Oct 2026 · Chart · CSVView →
-
AI BenchmarksJT3.5 and JT-VL-Chat tie for first on the MVBench video test at 74.1%; the top 15 span 23 points19 Oct 2026 · Chart · CSVView →
-
AI BenchmarksNo model reached the 74.4% human baseline on Video-MMMU; the best, Keye-VL-1.5-8B, scored 66%19 Oct 2026 · Chart · CSVView →
-
AI BenchmarksOn Video-MMMU's measure of learning from video, human experts gain 33.1% and the best model, GPT-4o, 15.6%19 Oct 2026 · Chart · CSVView →
-
AI BenchmarksGoogle's gemini-3-pro leads the Vision Arena at 1,288; Gemini models hold four of the top six places19 Oct 2026 · Chart · CSVView →
-
AI BenchmarksGemini 3.1 Pro Preview scored 88.2% on MMMU, 0.4 points short of the best human expert19 Oct 2026 · Chart · CSVView →
-
AI BenchmarksThe best of seven AI models read analogue clocks exactly right 22.6% of the time on ClockQA19 Oct 2026 · Chart · CSVView →
-
AI BenchmarksGPT-o1 answered 80% of CalendarQA date questions correctly; no other model of seven passed 47%19 Oct 2026 · Chart · CSVView →
-
AI BenchmarksThe best AI model read 50.6% of ClockBench clocks correctly in March 2026; humans manage about 90%19 Oct 2026 · Chart · CSVView →