How AI is applied across API Evangelist and APIs.io. Read my AI disclosure →
API Evangelist API Evangelist
Discovery
Learnings
Guidance
Toolbox
Alignment
API Evangelist LLC

Qwen2-VL-2B-Instruct.hef - 1 GB more between versions 5.1.1 and 5.2/5.2/5.4

calendar_today October 5, 2026 person @Leonid_Kogan domain hailo

I’ve noticed that the size and RAM usage of the model have increased by about 1 GB between versions 5.1.1 and 5.2. From the release notes, I can see the following: Load time: 9.71 → 7.21 seconds so about 26% faster. TTFT: 0.92 → 1.02 seconds so about 11% slower.

open_in_new Read original post