TeachObs: A Human-Validated Benchmark for Multimodal Teaching Observation and Model Evaluation
This benchmark provides a structured dataset for evaluating AI systems' ability to analyze teaching practices in classroom videos, which is an incremental step for researchers developing multimodal models for educational applications.
The researchers created TeachObs, a human-validated benchmark of 30 lesson videos (5,158 15-second scenes) annotated with 39 binary observation codes and expert lesson-level ratings. They evaluated five vision-capable frontier LLMs on this benchmark across three tracks, finding no single model consistently outperformed others, that adding a mid-frame inflated both true and false attributions, and that models over-rated procedurally clear lessons compared to experts.
Classroom videos contain observable teaching practices, but their pedagogical and visual signals are rarely organized in forms suitable for model evaluation. We present \textit{TeachObs}, a human-validated benchmark for multimodal teaching observation in classroom videos. \textit{TeachObs} includes 30 public lesson videos from eight countries divided into 5,158 fixed 15-second scenes. Seven researchers annotated each scene with 39 binary observation codes, covering 20 visual codes, such as gesture, board work, pointing, and visual materials, and 19 nonvisual codes, such as instruction, monitoring, questioning, feedback, and reflection. Gold segment labels are constructed using reliability- and prevalence-aware rules based on Krippendorff's alpha. In addition to segment-level labels, three expert raters produced lesson-level ratings and qualitative evaluations of instructional design, instructional delivery, learner response, learning materials, and lesson closure across the 30 lessons, with rater coverage detailed in the body. Using these two human reference layers, we evaluate five vision-capable frontier LLMs across three tracks - text-only segment coding, text + frame segment coding, and lesson-level coverage scored under an LLM-as-judge protocol - and find that no single model consistently outperforms others across all three tracks, that adding a mid-frame inflates both true and false attributions per scene, and that model evaluations over-rate procedurally clear lessons relative to expert raters. \textit{TeachObs} therefore supports both fine-grained annotation benchmarking and whole-lesson evaluation, showing where AI systems can assist classroom video analysis and where expert judgment remains necessary across varied subjects, classroom formats, and annotation difficulty levels.