Merge Gateway adds MiniMax H3-Image

AI gateways are expanding beyond text LLMs into multimodal generation as open-weight image models achieve production quality. Adding MiniMax's unified diffusion model to Merge Gateway gives developers centralized control over high-resolution generation alongside their existing LLM traffic.

• –MiniMax H3-Image combines text-to-image synthesis and multi-reference editing in a single checkpoint, removing the need for separate inpainting workflows.
• –Native 4K rendering in under 10 seconds bypasses secondary upscalers, substantially cutting inference latency for customer-facing creative tools.
• –Support for up to 16 reference images improves visual consistency across character design, branding, and multi-turn product image iteration.
• –Merge Gateway provides unified billing, spend caps, and failover routing for teams deploying open-weight vision models alongside commercial foundation APIs.