Banana 2.1 Out: Seedream Leads on Facial Consistency, Midjourney Still Unmatched in Aesthetics
hq4ai · x · 2026-10-07
With Nano Banana 2.1's release, an AI image blogger reviews the current landscape: for stylized aesthetics, Midjourney 8.2 remains in a class of its own; for precision needs (text covers, diagrams, info cards), OpenAI and Google models fit better.
On facial consistency (tested on an East Asian face "Yuri"): ByteDance's Seedream performs best, ChatGPT Images 2.5 is mediocre, Banana 2.1 is poor, and Midjourney doesn't offer it. The author notes no model yet nails high-demand facial consistency, though most real use cases like short dramas don't require it.
More from Multimodal
- LiFT loops a shared DiT to beat dense DiT-XL/2 with 60% fewer parameters on ImageNet — retr0jirachi · 2026-10-07
- E-commerce product showcase animation built with Opus 5.5 + Higgsfield — Tegadesigns · 2026-10-07
- Gemini Omni 1.1 Can Be Surprisingly Cohesive When Used Right, Demo Shows — Cold_Solder_ · 2026-10-07
- EmbeddingGemma 2 Demo: One Search Across Images, Video and Audio in Browser — tomaarsen · 2026-10-07
- Creator builds 4-minute AI sci-fi trailer MORO with Agent Two and AI-assisted sound design — LudovicCreator · 2026-10-07
- Solo game dev says GPT generated 1,200+ art assets for his airship game, 100x faster — MosskeepForest · 2026-10-07