Observed Signal · Jun 22, 2026 · Deprecation · Source: DEV Community · Impact: 4/5 · Sentiment: Negative
Google Imagen Models Shut Down; Migrate to Gemini Image
Google will shut down Firebase AI Logic Imagen image models on June 24, 2026 (imagen-3.0-generate-002, imagen-4.0-generate-001, imagen-4.0-ultra-generate-001, imagen-4.0-fast-generate-001). The Imagen mask-based editing capability (imagen-3.0-capability-001) is scheduled for shutdown on June 30, 2026 and currently has no direct Gemini Image replacement. Google’s recommended replacement is Gemini Image (public model gemini-2.5-flash-image, internal name “Nano Banana”). Migration requires renaming methods (generate_images → generate_content), swapping model identifiers, and updating response parsing; response structures differ and Gemini Image can return mixed text+image parts. Vertex AI legacy Imagen endpoints have an August 17 hard shutdown (Google recommends migrating before June 30). Gemini Image is faster but billed by output tokens; third-party inpainting services are suggested as fallbacks for mask-editing workflows.
Major platform (Google) is deprecating widely used image-generation endpoints with hard deadlines, requiring code and prompt migrations, creating a functional gap for mask-based editing, and affecting quotas and billing—this materially affects devs and product workflows that generate or edit images.
Track Google Signals & Market Shifts in Real-Time
Polaris7 autonomous intelligence agents track regulatory filings, primary sources, executive changes, and deal flow 24/7. Create your free Explorer workspace to monitor these entities.
Key Takeaways & Evidence Grounding
- Firebase AI Logic Imagen image models (imagen-3.0-generate-002, imagen-4.0-generate-001, imagen-4.0-ultra-generate-001, imagen-4.0-fast-generate-001) shut down on 2026-06-24.
- Imagen mask-based editing endpoint (imagen-3.0-capability-001) shuts down on 2026-06-30 and has no direct Gemini Image replacement.
- Google’s replacement image model is gemini-2.5-flash-image (Gemini Image); migration entails method rename (generate_images → generate_content), model-id change, and response-structure updates.
- Vertex AI legacy Imagen endpoints have a hard shutdown on 2026-08-17; Google recommends migrating before 2026-06-30 and checking quotas for gemini-2.5-flash-image.
- Gemini Image is reported faster (≈2–3s per 1024×1024 image) than Imagen 4 (≈4–6s) and uses per-output-token billing; SynthID watermark metadata format differs between Imagen and Gemini Image.
Connected Companies & Entities
2 Entities mappedOntology Mapping & Concepts
Related Market Signals & Shifts
Recent verified developments and strategic activity across this market segment.
Google Unveils Nano Banana 2 for Faster Image Generation
Google announced Nano Banana 2 (technically Gemini 3.1 Flash Image), an updated image-generation model that creates more realistic images faster than prior versions. Nano Banana 2 will become the default image model in the Gemini app across Fast, Thinking, and Pro modes, and will be used by default in Google Search (via Lens and AI Mode) across 141 countries, as well as in the Flow video-editing tool. The model supports resolutions from 512px to 4K, preserves character consistency for up to five characters and fidelity for up to 14 objects, and produces richer lighting, texture, and detail. Developers can preview the model via the Gemini API, Gemini CLI, Vertex API, AI Studio, and Antigravity. All generated images will carry Google’s SynthID watermark and will interoperate with C2PA Content Credentials; Google says SynthID has been used over 20 million times since its November launch.
Google expands Gemini with personalized AI image creation
Google has expanded personalized image-creation features in its Gemini app, enabling qualified users (initially in the U.S.) to generate highly personalized images by linking Personal Intelligence with apps like Gmail and Google Photos. The article describes Gemini Omni Flash — visible on Gemini's site — which supports 10-second video creation, native audio generation, videos created from up to five photos, and multi-step editing; access requires users to be 18+ and have a Google AI subscription. Google stresses that linking is opt-in and reversible but gives few details about whether linked user content is used to train models. The piece flags industry privacy and advertising implications because personal images and Gmail context could be leveraged, directly or indirectly, for personalization and ad targeting. The article also references other model names (Nano Banana 2, Gemini 3.5 Pro) and compares voluntary oversight practices to those of OpenAI.
Nano Banana 2: Faster AI Images with Gemini Knowledge
Google unveiled Nano Banana 2, an AI image-generation model that blends Gemini’s real-world knowledge with capabilities to render visuals directly from text, enabling faster creation of infographics and marketing mockups. The update adds improved image-generation performance categories, including character consistency up to five and object consistency up to 14, better instruction-following, and direct control over asset specs and templates. It supports new aspect ratios (4:1, 1:4, 8:1, 1:8) and showcases enhanced lighting and textures. Rollout spans multiple Google environments: Nano Banana 2 is available in the Gemini App for AI Pro/Ultra, integrated in AI Mode, Lens, browser, and the app; Preview in AI Studio and Vertex AI via the Gemini API; and for Flow users as a base model with no credits required. An Ads feature allows marketers to generate asset suggestions for campaigns. Pricing is 60 USD per million tokens, with per-image costs ranging from $0.045 (512px) to $0.151 (4K). SynthID and interoperable C2PA Content Credentials are included to indicate AI usage; Pratham Kumar, an AI researcher, notes potential handwriting replication concerns.
Track Real-Time Market Signals & Shifts
Set up custom watchlists to receive automated, evidence-grounded executive digests whenever material signals or shifts occur across your tracked landscape.