Test shows Qwen3.8-27B achieves high precision in long video timestamping
vanstriendaniel · x · 2026-08-15
A user tested the long video understanding capabilities of Qwen3.8-27B. In a single request, the model processed an 11-minute film from 1935, identifying 96 timestamped events and quoting on-screen text verbatim. Verification by extracting frames at the claimed timestamps showed an accuracy of within 2 seconds across the entire film, demonstrating strong performance in temporal reasoning and detail capture.
Related event: Qwen Model Accurately Analyzes 11-Minute Long Video(2 posts)→
More from Models
- Anthropic's internal model scores leak: new model may surpass Mythos 5 — scaling01 · 2026-08-15
- Qwen 3.8 27b Rivals Opus 4.6? Local Models Challenge Frontier AI, Reddit Debates — Odd_Tumbleweed574 · 2026-08-15
- FrontierSWE Rankings: GLM 5.3 Takes #2, Grok 4.6 #3, Claude Fable 5 Still Top — dejavucoder · 2026-08-15
- OpenAI's o3 Model Non-Functional for Paying Users; User Organizes Complaint Campaign — Aine_123 · 2026-08-15
- Arcee Adds deepseek-v4-pro-0813 to Open Models API — latkins · 2026-08-15
- Three major AI releases: Qwen3.8-27B, GLM-5.3, DeepSeek V4-Pro — johnseach · 2026-08-15