Models · Model provider
MiniMax
MiniMax develops language, image, video and audio models and provides a hosted developer API.
+ FollowBuilder facts
- Language models. MiniMax M3 and M3.1-Flash support multimodal input and a one-million-token context. Flash exposes adjustable thinking depth. Source ↗
- Direct API access. Direct API calls require a MiniMax account and an API key, with pay-as-you-go billing or a separate M Plan subscription. Source ↗
- Local deployment. The experimental SGLang reference deployment for M3.1-Flash uses four NVIDIA B300 GPUs. The documented H200 configurations cap context at 140,000 tokens. Source ↗
- Music API availability. The paid music and lyrics APIs closed to new users on August 20, 2026. Existing paying users retain access, and MiniMax Music 3 weights remain downloadable. Source ↗
Sources checked
Publisher update ·
MiniMax releases H3-Image ↗MiniMax released weights and direct API access for text-to-image generation and edits with up to 16 reference images.
Change timeline
0 storiesNo verified releases in this timeline yet.