H3-Image Text-to-Image Eval: Leads Peers on Layout, Trails GPT Image 2.5 by Up to 269 Elo
ArtificialAnlys · x · 2026-11-14
Per Artificial Analysis's Image Arena taxonomy (9 capabilities, 10 use cases for text-to-image):
- Human Anatomy & Knowledge: H3-Image comes closest to the leaders, yet still trails GPT Image 2.5 Sunburst (max) by 149 Elo in Anatomy and 160 in Knowledge.
- Text Rendering: 269 Elo behind GPT Image 2.6 Sunburst (max).
- Layout: clearly ahead of same-tier Krea 2 Large and MAI-Image-2.5-Flash.
- Text Rendering vs peers: clearly behind Qwen-Image-3.0 and Qwen-Image-3.0-Pro.
Text-to-image is H3-Image's weaker board: every gap here is wider than any of its gaps in image editing.
Related event: H3-Image Leads in Some Categories but Trails GPT Image in Arena Ratings(3 posts)→
More from Multimodal
- Full-duplex speech model Haru-4B trends on Hugging Face — higurashi-ai · 2026-11-14
- ByteDance releases Seedance 2.5 Flash video model at 35% off — videorouter · 2026-11-14
- OpenAI Images 2.6 tops DesignArena and Image Arena at 4s per image, same price — thursdai_pod · 2026-11-14
- MiniMax H3 Camera & Motion Control Tested on a RTX 4070 with Only 8GB VRAM — big-boss_97 · 2026-11-14
- 3D scene gen comparison: Inkling 2 Small cinematic, Kimi K3 bland — LiTianleli · 2026-11-14
- Anthology episode '2000' made entirely with AI tool Cuelist — Uncanny_Harry · 2026-11-14