SOTA Temporal Video Understanding on TVBench (test) and PapersWithCode

51.4Accuracy

Structured Reward Reinforcement Learning

Updated 16d ago

Evaluation Results

Method	Links
Structured Reward Reinforcement Learning 2026.04		51.4
Qwen3-VL-32B 2026.04		51.3
Season 2026.04		49.5
ArrowRL 2026.04		49.4
Structured Reward Reinforcement Learning 2026.04		48.6
Structured Reward Reinforcement Learning 2026.04		48
Season 2026.04		47.7
Gemini 1.5 Pro 2026.04		47.6
Structured Reward Reinforcement Learning 2026.04		47.1
Qwen3-VL-8B 2026.04		47
DINO-HEAL 2026.04		46.9
Season 2026.04		46.8
Qwen2.5-VL-7B 2026.04		46.7
TCD 2026.04		46.1
Season 2026.04		43.2
Season 2026.04		43
Structured Reward Reinforcement Learning 2026.04		43
TCD 2026.04		42.9
DINO-HEAL 2026.04		42.9
LLaVA-OV-7B 2026.04		42.8
Qwen3-VL-4B 2026.04		41.7
GPT-4o 2026.04		39.9