MiniMax H3-Image Takes #1 Open-Weights Spot on Both AA Image Leaderboards
ArtificialAnlys · x · 2026-11-14
Artificial Analysis reports that MiniMax H3-Image is the new #1 open-weights image model on both AA-Image-Editing v2.1 and AA-Image-T2I v2.1. Overall it ranks #13 in image editing and #25 in text-to-image; the API-only faster variant H3-Image Highspeed ranks #14 and #24.
Related event: MiniMax H3-Image Tops Both Open-Weight Image Leaderboards(2 posts)→
More from Multimodal
- Ask Claude to reverse-engineer prompts from videos — prompt hoarding is dead — AIandDesign · 2026-11-14
- HeyGen agent auto-produces a full nature documentary with Video 1.0 via MCP — HeyGen · 2026-11-14
- Fully local meme person-swap workflow: MiniMax H3 + Pika in ComfyUI, runs on 8GB VRAM — AuzBoss · 2026-11-14
- Stunning AI-generated floral portraits showcase how far image models have come — gizakdag · 2026-11-14
- Removing dimension callouts from product photos with Qwen Edit + Kontext — dashed lines remain hard — Sarsippius3 · 2026-11-14
- AI-generated 'Are you firing me?' short video hits Reddit — Darri3D · 2026-11-14