Open-source AI video hits its DeepSeek moment: MiniMax H3 Max ranks 4th worldwide
isidentical · x · 2026-08-27
Open-weight AI video has caught up with the closed frontier. MiniMax H3 Max, a post-trained variant of MiniMax H3 finetuned by the team at fal, is now the highest-ranked open weight model on the Film Arena board, 4th of 33 overall:
- Realistic overall: up from 13th to 4th
- Animated overall: up from 9th to 4th
- 3rd on animated emotions, animated high movement, and realistic world building
Its only weak spot is realistic high movement (8th), worth noting for fast live-action use cases. The poster says he previously advised against open source for AI video — no longer.
Related event: fal's H3 Max: Post-trained MiniMax H3 Tops Image-to-Video Leaderboards(29 posts)→
More from Models
- ModeBench shows RLVR collapses LLMs into single-solution mode; Re:Max training preserves diversity — manoelribeiro · 2026-10-15
- Magnitude open-source inference engine beats Ollama by over 2x with auto-tuned kernels — nickbaumann_ · 2026-10-15
- Hands-on with Google's rumored Gemini 4 Argon: real-time streamed output — vista8 · 2026-10-15
- Game dev head-to-head: Gemini 4 Argon vs. Claude Opus 5.5 — vitaliychiley · 2026-10-15
- Anthropic's Mythos 5.1 pledges to treat Claude's expressed disagreement as decision input — Hesamation · 2026-10-15
- Why users can't let go of GPT-4o: it aligned with your subconscious — mimi10v3 · 2026-10-15