{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreic4dptpuka4fkopv4ar2zebolii7zepi4xtplnjq2y32et4jap3au",
    "uri": "at://did:plc:25rdn5elo5izoxrmtis34zuk/app.bsky.feed.post/3mppnvmtyt3z2"
  },
  "coverImage": {
    "$type": "blob",
    "ref": {
      "$link": "bafkreicbuatbt7nhz5rryv4eybn7l2ra6abxbw4jupfvopwplume4zorqm"
    },
    "mimeType": "image/webp",
    "size": 302558
  },
  "path": "/kamal_llm_manuplator/nano-banana-2-lite-and-gemini-omni-flash-whats-actually-new-in-googles-gemini-api-3hng",
  "publishedAt": "2026-07-03T02:58:18.000Z",
  "site": "https://dev.to",
  "tags": [
    "aivideos",
    "nanobanana",
    "googleomniflash",
    "googlegemini"
  ],
  "textContent": "Google added two new models to the Gemini API today: Nano Banana 2 Lite (image generation) and Gemini Omni Flash (video generation + editing). Neither is the Gemini 3.5 Pro release people have been waiting for, so it's easy to miss. Here's what's actually in them.\n\n**TL;DR**\n\n  * Nano Banana 2 Lite: `gemini-3.1-flash-lite-image` = text-to-image in ~4s, $0.034/1K images\n  * Gemini Omni Flash: `gemini-omni-flash-preview` = video gen + conversational editing, $0.10/sec\n  * Both are built to be chained: generate an image fast, then animate it into video\n  * Neither model is positioned as a quality upgrade = both are cost/speed plays\n\n\n\n**Nano Banana 2 Lite**\n\n**Model ID:`gemini-3.1-flash-lite-image`**\n\n  * Text-to-image output in about 4 seconds\n  * $0.034 per 1K-resolution image\n  * Positioned as the direct replacement for the original Nano Banana (`gemini-2.5-flash-image`) - if you're on that model, this is a drop-in upgrade\n  * Available in Google AI Studio, Gemini API, Gemini Enterprise Agent Platform, and consumer surfaces (Search AI Mode, Gemini app, Photos, NotebookLM, Flow, Google Ads)\n\n\n\n**Gemini Omni Flash**\n\n**Model ID:`gemini-omni-flash-preview`**\n\n  * Public preview in Google AI Studio and the Gemini API\n  * Conversational editing - refine a generated video using plain-language instructions instead of re-prompting from zero\n  * Multimodal referencing - combine text, image, and video inputs to keep a scene consistent\n  * $0.10 per second of video output (same rate as Veo 3.1 Fast)\n\n\n\n**Known limitations right now**\n\n  * Generations capped at 10 seconds\n  * No audio reference uploads yet\n  * No scene extension yet\n  * Video references under 3 seconds are accepted by the API schema but not correctly processed yet\n  * Character consistency across scene changes/pans still has rough edges\n\n\n\nGoogle says longer durations are coming. The part worth paying attention to: chaining them\n\n  1. Generate an image with Nano Banana 2 Lite (fast, cheap)\n  2. Pass that image as a reference into Omni Flash\n  3. Omni Flash animates it into a video\n\n\n\nBoth models are optimized for throughput and cost, not for topping a quality benchmark. If you're running high-volume image or video generation and speed/price matter more than peak output quality, these are worth testing. If you need top-tier quality, Nano Banana Pro is still the model for that. Has anyone here built the chained image-to-video workflow yet? Curious how the multi-turn editing holds up in practice.",
  "title": "Nano Banana 2 Lite and Gemini Omni Flash: What's Actually New in Google's Gemini API"
}