[Core Function] Gemini 3.8 Flash Lite TTS is Google's text-to-speech model for low-latency speech generation. [Strengths] Single-speaker synthesis, two-speaker dialogue, 101 languages, per-turn style control, and WAV output (24 kHz mono PCM). [Best For] Bulk narration, voice-agent responses, read-aloud features, and multilingual dialogue. [Limitations] Do NOT use this for streaming, voice creation, music generation, or audio, image, or video input. Dialogue requires exactly two speakers with prebuilt or extended-library voices; designed or replicated voices are limited to single-speaker synthesis. Combined text and style must stay within 8,192 Unicode characters. [Routing] Choose Flash Lite for latency and cost efficiency; choose Gemini 3.8 Flash TTS for expressive delivery and broader language coverage.
[Core Function] Gemini 3.8 Flash TTS is Google's controllable text-to-speech model, currently in Preview. [Strengths] Single-speaker and two-speaker dialogue, preset voices, 130 languages via language_code, and expressive delivery through turn-level style and inline vocal tags. Output is WAV (24 kHz mono PCM). [Best For] Voiceovers, audiobook narration, multilingual speech, podcast-style scripts, and two-person dialogue. [Limitations] Do NOT use this for streaming, voice creation, lip-syncing existing video, music or sound-effect generation, image or audio inputs, MP3/OGG export, or API-level speed, volume, encoding, or sample-rate controls. Dialogue requires exactly two speakers with preset voices. Combined text and style must stay within 8,192 Unicode characters. [Routing] Route here for expressive Gemini TTS with per-turn style control; for the prompt-based request format use Gemini 3.1 Flash TTS; for video with native audio use Veo.
[Core Function] Grok Imagine Video 1.5 Lite I2V animates a single starting image into a video with native audio. [Strengths] It uses the input image as the first frame, keeps the output aspect ratio aligned with that image, and adds lip-synced speech. [Best For] Highly recommended for: animating a photo or illustration into a short clip when the Lite backbone is enough. [Limitations] Do NOT use this model for text-only generation, reference images, voice references, pinned frames, video editing, or video extension. It requires one starting image. Duration is 1-15 seconds and resolution is 480p, 720p, or 1080p. 1080p is upscaled from 720p, not native 1080p. The output aspect ratio follows the input image. [Routing] Use this model when the user provides one starting image and wants the Lite backbone. For text-only video, use Grok Imagine Video 1.5 Lite. If the user needs reference images, voices, pinned frames, or native 1080p, use the full Grok Imagine Video 1.5 models.
[Core Function] Grok Imagine Video 1.5 Lite generates a short video with native audio from a text prompt. [Strengths] It produces lip-synced speech at the lowest price in the Grok Imagine video family, with controllable duration, aspect ratio, and resolution. [Best For] Highly recommended for: simple social clips, animated concepts, and text-only drafts that do not need reference images or pinned frames. [Limitations] Do NOT use this model when a starting image is provided, or for reference images, voice references, first or last frames, keyframes, video editing, or video extension. Duration is 1-15 seconds and resolution is 480p, 720p, or 1080p. 1080p is upscaled from 720p, not native 1080p. [Routing] Use this model for text-only video on the Lite backbone. If the user provides one starting image, use Grok Imagine Video 1.5 Lite I2V. If the user needs reference images, voices, pinned frames, or native 1080p, use the full Grok Imagine Video 1.5 models.
[Core Function] Vidu Q4 R2V is the preview reference-to-video model that composes a new 3-16 second clip from 1 to 15 reference images while keeping subject appearance consistent. [Strengths] It supports 540p through 4K, aspect ratios including 3:4 and 4:3, synchronized audio, and up to three reference audio clips. [Best For] Highly recommended for: multi-image character consistency, vertical 3:4 portraits, and scenes that combine reference images with spoken audio. [Limitations] Do NOT use this model to animate one image exactly as the first frame; use Vidu Q4 I2V for that. Do NOT omit the prompt. Do NOT send more than 15 images or more than 3 audio files. Subject-library invocation is not supported. [Routing] Use this model when the user supplies reference images and a prompt for a newly composed shot. Use Vidu Q4 I2V for a single start frame.
[Core Function] Vidu Q4 I2V is the preview flagship image-to-video model that animates one starting frame into a 3-16 second clip with synchronized audio and automatic camera switching. [Strengths] It keeps the supplied image as the opening frame, supports 540p through 4K, and can generate dialogue and sound effects in the same pass. [Best For] Highly recommended for: animating a key still, product or character shots that must begin from a specific image, and short clips that need native sound. [Limitations] Do NOT send more than one image. Do NOT use this model for multi-image subject reference or a chosen output aspect ratio; the frame follows the first image. Do NOT use a duration under 3 seconds or over 16 seconds. Recommended-prompt mode is on by default, so a prompt is optional. [Routing] Use this model when the user provides a single start frame. Use Vidu Q4 R2V when the user provides reference images and wants a newly composed shot with a chosen aspect ratio.
[Core Function] Nano Banana 2.1 Edit is Google's fast image-editing and video-to-image model, exposed as google/nano-banana-2.1-edit. [Strengths] It edits up to 14 reference images, or one HTTPS video plus up to 10 images, at 1K, 2K, or 4K, with improved character and multi-reference consistency. thinkingLevel defaults to medium. Optional systemInstruction is omitted when blank. enableWebSearch turns on Google Search grounding (web and image search) only when true. [Best For] Prompted restyling, multi-image product and character edits, and turning a video into a still. [Limitations] Do NOT use this model for text-only generation, 512-pixel output, or audio generation. At least one of image or video is required. seed, temperature, topP, topK, and logprobs are rejected. [Routing] Use this endpoint when the user provides images or a video and wants Nano Banana 2.1. Use google/nano-banana-2-edit for the previous Gemini 3.1 Flash Image editor, including 512 output. Use google/nano-banana-2.1 when there is no visual input. Use google/nano-banana-pro-edit for the highest-fidelity still edits that do not need video input.
[Core Function] Nano Banana 2.1 is Google's fast text-to-image model on Gemini 3.6 Flash, exposed as google/nano-banana-2.1. [Strengths] It keeps Flash-level speed at 1K, 2K, and 4K, with stronger visual quality, text rendering, and real-world knowledge than Nano Banana 2. thinkingLevel trades latency for quality and defaults to medium. Optional systemInstruction is omitted when blank. enableWebSearch turns on Google Search grounding (web and image search) only when true. [Best For] High-volume image production, posters and diagrams with readable text, localized text, and rapid iteration on new text-to-image work. [Limitations] Do NOT use this model for image or video editing, 512-pixel output, audio input, or reference images. This endpoint accepts a prompt only. seed, temperature, topP, topK, and logprobs are rejected. [Routing] Use google/nano-banana-2.1 for new text-to-image work. Use google/nano-banana-2 when the caller needs the 512 resolution of Gemini 3.1 Flash Image. Use google/nano-banana-2.1-edit when the user supplies images or a video. Use google/nano-banana-pro when the user wants the highest-end professional stills.
[Core Function] MiniMax H3 Max V2V generates video guided by one or more reference videos plus a text prompt. [Strengths] It accepts up to 3 reference videos with optional reference images and audios, 5-15 second duration, 480P or 768P output, optional aspect ratio defaulting to 16:9, and optional prompt_expansion_mode (disabled, balanced, or quality). [Best For] Highly recommended for: motion remix, style transfer from video clips, keeping subject motion while changing the scene description, and multi-clip reference guidance. [Limitations] Do NOT use this model for text-only generation, first or last frame transitions, or if you need 2K or a 4 second clip. prompt, duration, resolution, and reference_videos are required. Do NOT send first_frame_image or last_frame_image on this endpoint. [Routing] Choose MiniMax H3 Max V2V when the user provides reference videos. For reference stills only, use MiniMax H3 Max I2V. For start or end frames, use MiniMax H3 Max FL2V. For text only, use MiniMax H3 Max T2V. For 2K or 4 second clips, use MiniMax H3 V2V.