Google Shuts Down Imagen 4 Endpoints in the Gemini Developer API on August 17

Google's Gemini Developer API documentation lists August 17, 2026 as the shutdown date for three Imagen 4 model identifiers: imagen-4.0-generate-001, imagen-4.0-ultra-generate-001 and imagen-4.0-fast-generate-001. Google's documentation warns that the models are deprecated and instructs developers to migrate to a newer image generation model to avoid service disruption. Google's current image guidance points developers toward its newer Gemini image family for all new generation work, positioning Gemini 3.1 Flash Image as the general-purpose option balancing speed, quality and cost. The newer model accepts text, image and video context and supports conversational editing, search grounding, video-to-image workflows and output up to 4K resolution. The migration is not a straight substitution. Applications that call the older generate_images() method must be reworked, and behavior such as generating multiple variants in a single call has to be restructured. Google does not publish an exact cutoff hour, which leaves teams that have not tested replacement prompts exposed to output changes or failures. Pricing also shifts. Imagen 4 used flat per-image rates of roughly $0.02 for Fast, $0.04 for Standard and $0.06 for Ultra. Gemini 3.1 Flash Image is billed by output tokens, working out to about $0.045 per image at 0.5K, $0.067 at 1K, $0.101 at 2K and $0.151 at 4K, with input, text-output and search-grounding charges potentially applying on top. For standard-resolution work that represents an increase, although batch processing can bring the effective per-image cost back down. Google Cloud's Vertex AI platform follows a separate model lifecycle; customers there had already been advised to update their Imagen 4 integrations earlier in the year, so the August 17 deadline applies to the Gemini Developer API schedule. Developers migrating are advised to test representative prompts covering text rendering, aspect ratios, safety blocks, latency and image consistency before shifting production traffic.