MiniMax launches M3. 1-Flash-Preview AI model for developers

AI developer MiniMax has launched its new text model, MiniMax-M3. 1-Flash-Preview. Designed specifically for software developers, the model is integrated into the MiniMax Code environment to assist with routine programming tasks, including bug localization, code segment implementation, and testing.

The model is optimized for speed and stability to facilitate rapid debugging. It features a 1-million-token context window and supports multimodal inputs, including text, images, and video. Users can adjust the depth of reasoning through an ‘effort’ parameter with five levels ranging from low to max.

Currently, access to the M3. 1-Flash-Preview is limited to the MiniMax Code product and the subscription-based Token Plan. The model supports both Anthropic-compatible and OpenAI-compatible APIs, allowing developers to integrate it into existing SDKs by updating the base URL and model name. To encourage adoption, MiniMax has announced a bonus program offering double free points for daily check-ins through October 7, 2026.