Atlas / Open-weight and challenger labs / MiniMax
MiniMax
AI Market Atlas · by Pro Pluribus
China
M3 open model, plus Hailuo video and consumer apps.
What it offers
Main offerings
- MiniMax M3
- Open weights, 1M-token context, image and video input.
- Hailuo 3.0 · Talkie
- Hailuo 3.0 video model (July; open weights for H3-Base); companion app.
Recent news
- MiniMax open-sources H3 video model weightsOpen weights for a frontier-grade video model put pressure on closed rivals such as Kling, Veo and Runway.ComfyUI Wiki ↗
- MiniMax launches Hailuo 3.0 omni-modal video modelReportedly #1 in video editing on Artificial Analysis; unified multimodal input is a structural change.LLM Stats ↗
- MiniMax releases M3 with a 1M-token context and native multimodal inputFrontier-level coding and agent performance from a lab listed in Hong KongWikipedia ↗
- MiniMax M2.5 scores 80.2% on SWE-Bench Verified at about $1 an hourSet the benchmark for cheap frontier-level agent models soon after the Hong Kong IPOMedium (Maxime Labonne) ↗
This map is a by-product of the work. Keeping track of what each of these actually does, and which ones are worth your time, is part of how I advise founder-led companies on AI. If you want that applied to your business, the first conversation is free.
Book a 30-minute conversation →Compiled from public sources by Pro Pluribus; updated 10 Nov 2026. General information, not advice. Company names and logos belong to their owners and are used to identify them, without affiliation or endorsement. Terms