A MiniMax lançou o H3 em 31 de julho. A combinação chama atenção: preço comercial de API e plano de abrir os pesos do modelo.
O H3 aceita texto, imagens, vídeo e áudio no mesmo contexto. Gera até 2K e 15 segundos, com áudio estéreo nativo.
A abertura ainda depende dos arquivos
“We plan to open up the model weights in the coming days, subject to applicable laws and regulations.”
Developers still need the license, commercial-use terms, hardware guidance and a technical report before treating H3 as a deployable open model.
O preço muda a comparação
Shanghai Securities News, The Beijing News and 21st Century Business Herald reported H3 at RMB 0.8 per second for 2K generation, about one-third of comparable flagship video models. A 15-second 2K clip is roughly RMB 12 before platform packaging.
A aposta técnica é contexto unificado
MiniMax describes H3 through Contextual Omni Representation, H3-VAE, H3-Omni Transformer and In-Context Regeneration. The model handles video, audio, images and text as one creative context.
Most source material requires about 100,000 tokens of inference before being distilled to roughly 4,000 tokens on average. H3-VAE provides a 4x gain in effective sequence length, and workload separation lifted training throughput by nearly 30%.
The next checks are whether weights appear, whether the license works for companies, and whether the community can reproduce usable text, motion and audio control on non-official hardware.
Fontes: blog oficial de pesquisa da MiniMax, Shanghai Securities News, The Beijing News, CocoLoop, 21st Century Business Herald; site da MiniMax para verificar 2K/15 segundos, estéreo nativo, compromisso de pesos abertos, preço de RMB 0,8 por segundo, comparação de 768p e ganho de quase 30% no throughput de treino.