GPT-Image-2.6-Flare: Fast image generation variant, optimized for low latency. Supports text-to-image, image-to-image and reference locking, ideal for high-concurrency rapid rendering.