Qwen-Image-2.1 vs SenseNova U1.5 Lite: 5 same-prompt image edit tests show an instruction-following gap
Wild_Artichoke7648 · reddit · 2026-09-24
The author ran Qwen-Image-2.1 and SenseNova U1.5 Lite on five editing tasks with identical inputs and prompts.
- Marker-guided editing: Qwen's output looks lush but leaves annotation lines; SenseNova removes all marks and adds every requested element.
- Text replacement: Qwen has stronger typography but misses several replacements; SenseNova completes them with minor garbled footer text.
- Multi-reference poster: Qwen keeps the band photo sharp but loses titles and layout; SenseNova preserves structure and restyles members to match.
- Multi-input food scene: Qwen has natural lighting but drops the curry udon; SenseNova follows placement and labels better.
- Six-person portrait: both convincing, neither preserves every face.
Key takeaway: Qwen's weakness is instruction adherence, not image quality—it often misses explicit requirements in complex prompts. SenseNova is less dramatic visually but more consistent with multi-part constraints. Parameter counts aren't directly comparable: Qwen's 7B is the visual generation component (paired with a Qwen3-VL 8B encoder), while SenseNova U1.5 Lite uses a unified encoder-free, VAE-free 8B-MoT architecture.
More from Models
- Claude Opus 5.5 Designs a Buildable LEGO Microduck: 1,113 Real Parts, Zero Collisions — soleio · 2026-09-24
- Sarvam Vision 2.1 launches with 94.97% on OmniDocBench and OCR for 22 Indian languages — itsOmSarraf_ · 2026-09-24
- Why Persimmon builds human-behavior models on NVIDIA Nemotron 3 Ultra base instead of an aligned assistant — CShorten30 · 2026-09-24
- Jev-as-a-Judge: cheap judge cascade keeps 99% of GPT-6 accuracy at 57% of the cost — dair_ai · 2026-09-24
- Odyssey Launches Agora-2, a Multi-Agent World Model Simulating Up to 20 Humans and Agents in Real Time — damianplayer · 2026-09-24
- Even an Elon-endorsed take: inferior models in better products now beat better models in worse apps — elonmusk · 2026-09-24