Models.
Nano Banana 2.1
Nano Banana 2.1 is Google's efficient workhorse image model built on Gemini 3.6 Flash. It combines native 1K to 4K output, clear and readable in-image text, and subject consistency with up to 10 reference images, making it well suited to high-volume production of ecommerce posters, product variations, infographics, and character visuals.
GPT-6.1 Sol
gpt-6.1-sol is OpenAI’s reasoning model in the GPT-6 family for complex coding and agent development, using long code context to trace bugs, refactor projects, and plan tool calls across multi-step engineering tasks.
Claude Sonnet 5.5
claude-sonnet-5-5 is Anthropic’s Sonnet-family model for fast everyday coding and knowledge work, combining reasoning, long-context analysis, and tool use to implement features, diagnose bugs, and prepare technical documentation.
GPT-6 Sol
gpt-6-sol is OpenAI’s GPT-6 reasoning model for complex coding and AI agents, combining long-context code analysis, debugging, image understanding, function calling, and structured outputs.
GPT-6 Luna
gpt-6-luna is OpenAI’s efficient reasoning model in the GPT-6 family, built for focused, high-volume work, with long-context understanding for customer support, document summaries, information extraction, and targeted coding tasks.
Claude Opus 5.5
claude-opus-5-5 is Anthropic's Opus model for complex coding and sustained agent workflows, bringing codebase reasoning, cross-file debugging and visual understanding to migrations, refactors and professional deliverables.
Grok 4.7
grok-4.7 is xAI’s Grok-family model for coding and professional knowledge work, combining sustained multi-step reasoning, stronger long-context management, and careful self-verification to help developers debug code, plan cross-file refactors, and analyze extensive technical documents.
Qwen Image 2.1
Qwen Image 2.1 is a lightweight, efficient, unified image model from Alibaba's Qwen team, built on a 7B single-stream DiT architecture. It combines text-to-image generation and multimodal editing, native RGBA output with transparent backgrounds, controlled blending of up to 10 reference images, and clear Chinese typography. It is suited to white-background product images, virtual try-on, advertising posters, and continuous storyboards.
Seedream 5.0 Flash
Seedream 5.0 Flash from ByteDance is a workhorse image model focused on speed and value. It combines responses in under 30 seconds, native 1K to 2K quality, clean text rendering, and blending of up to 10 reference images for batch production of game sketches, marketing posters, and commercial visuals.
DeepSeek V4.1 Flash
deepseek-flash is DeepSeek’s multimodal V4.1 MoE model for coding agents, combining a million-token context and native image understanding to connect code, screenshots and technical specifications throughout debugging, refactoring and tool-driven development.
GPT Image 2.5
OpenAI GPT Image 2.5 Flare and Sunburst. Flare is the everyday default with up to 50% lower latency than GPT Image 2; Sunburst holds finer detail for precision edits, transparent cutouts, and 1K to 4K output.
GPT-6 Astra
gpt-6-astra is OpenAI’s flagship reasoning model in the GPT-6 family, combining sustained reasoning, long-context understanding, and text and image input for complex coding, technical research, and professional document creation.
Gemini 3.8 Flash
Google's gemini-3.8-flash brings sustained reasoning to software engineering and coding agents, combining a 1M-token input window with image, audio, video, and PDF understanding to develop features, diagnose failures, and analyze large codebases.
Claude Fable 5.1
claude-fable-5-1 is Anthropic's Fable model for complex coding and sustained reasoning, combining a million-token context with image understanding to help agents coordinate multi-file development, investigate difficult defects and deliver clear research findings.
Gemini Omni 1.1 Flash
Gemini Omni 1.1 Flash turns text and multimodal assets into 4–10 second videos, with first and last frame interpolation, multiple image references, and motion guidance. It produces coherent, natural movement while preserving subject identity and environmental lighting.
MiniMax H3 Max
MiniMax H3 Max turns text prompts and images into high-fidelity video with fast generation, strong prompt adherence, and native synchronized stereo sound. It anchors first and last frame compositions and preserves character continuity, with dynamic lighting guided by camera movement descriptions.
MiniMax H3 Max Turbo
MiniMax H3 Max Turbo is a distilled, accelerated version of the flagship MiniMax video model, with twice the generation throughput at 50% lower cost. It supports text and first and last frames with native synchronized stereo sound, for frequent iteration, social shorts, and batch commercial production.
Wan 3.0 Video Prime
Wan 3.0 Video Prime turns text, images, and omnimodal references into HD videos up to 30 seconds long, with physical simulation and native synchronized audio. It keeps subjects and compositions stable through complex motion while supporting broad, natural camera and subject movement.
Gemini 3.7 Flash
gemini-3.7-flash is a previous-generation model in Google’s Gemini 3 family for coding and agent workflows, using million-token input context to analyze repositories, fix bugs, and develop web apps.
Grok 4.6
grok-4.6 is SpaceXAI’s Grok-family model for sustained coding and knowledge work, combining multi-step reasoning, codebase analysis, and self-checking to turn complex engineering tasks into tested changes, interactive application code, and clear research deliverables.
Pricing to be confirmed
Grok Imagine Image 2.0
Grok Imagine Image 2.0 is xAI’s flagship image model for the highest fidelity in the Grok Imagine visual family. It combines crisp English typography, consistent subjects guided by up to 3 reference images, and realistic lighting and textures, making it well suited to demanding commercial posters, e-commerce product images and brand visuals.
Wan 3.0
Wan 3.0 turns text, images, and omnimodal assets into coherent videos up to 30 seconds long, with native synchronized stereo audio and adaptive aspect ratios. It preserves subjects and realistic physics through large movements, with detailed lighting and expressive camera work.
Seedance 2.5
ByteDance Seed's Seedance 2.5 is a long-form multimodal video model built for coherent 4–30 second scenes, first-and-last-frame direction, dense image/video/audio references, and synchronized sound in ads, previs, and narrative clips.
Hailuo 03
MiniMax's Hailuo 03 is a native 2K video model for controlled 5–15 second clips from text, first and last frames, or image, video, and audio references, suited to cinematic shots, continuity work, transitions, and previs.
claude-opus-5
Anthropic's flagship Opus 5 model prioritizes deep decomposition, long-horizon execution, self-verification, and difficult coding or professional work when completion quality matters more than latency and token cost.
FLUX 3
Black Forest Labs' FLUX 3 video family supports prompt, source-image, first-and-last-frame, source-video extension, and positioned-keyframe workflows with native audio.
Qwen Image 3.0
Alibaba Qwen Team's image model for information-rich generation and reference-guided editing with long prompts, fine text, multilingual copy, and 1K or 2K output.
Kimi K3
Moonshot AI's Kimi K3 is a 2.8T sparse-MoE flagship with a 1M-token context window for long-horizon coding and knowledge work where sustained depth matters more than minimum latency or token cost.
GPT-5.6
OpenAI GPT-5.6 family with Luna for efficient workloads, Terra for balanced professional reasoning, and Sol for maximum coding and agentic capability.
Seedream 5.0 Pro
ByteDance Seed's Seedream 5.0 Pro is a professional image family for turning complex briefs into structured 1K or 2K visuals and making reference-guided revisions, suited to product explainers, editorial layouts, posters, diagrams, and presentation assets.
Omni Flash
Google Omni Flash combines prompt, image-reference, and one-video-guided workflows for multimodal short-video generation and revision, with 720p, 1080p, and 4K APIDot tiers.
Nano Banana 2 Lite
Google DeepMind's Nano Banana 2 Lite is the efficiency-focused Gemini 3.1 Flash-Lite Image family for rapid 1K generation and reference-guided editing, suited to storyboards, editorial illustrations, learning visuals, social assets, and early design directions.
claude-sonnet-5
Anthropic's balanced model for agentic coding, multi-step tool workflows, and long-context knowledge work through Messages or OpenAI-compatible chat.
Seedance 2.0 Mini
Seedance 2.0 Mini is a lower-cost video model for high-frequency short clips, prompt tests, reference-video motion transfer, and reference-guided 480p or 720p drafts.
Happy Horse 1.1
Happy Horse 1.1 is Alibaba's updated short-video model for text-to-video, single-image animation, and reference-guided scene generation, suited to teams that need stable motion, clear visual continuity, and 720p or 1080p output.
Grok Imagine Video 1.5
Grok Imagine Video 1.5 is xAI's image-to-video model for animating one reference image into short 480p or 720p clips with prompt-guided motion and duration control.
Claude Opus 4.8
Anthropic's Claude Opus 4.8 is a premium Opus model for 1M-token long-context reasoning, long-horizon agents, complex coding, adaptive thinking, tool-use planning, and high-value professional analysis.
Gemini 3.5 Flash
Google's Gemini 3.5 Flash is a stable Flash model for production traffic, with 1M-token long-context input, fast text reasoning, coding iteration, and agentic workflows for support assistants, knowledge-base Q&A, product copilots, and streaming chat interfaces.
Grok Imagine Image Quality
Grok Imagine Image Quality is xAI's higher-fidelity image model for production-ready text-to-image and multi-reference editing, with sharper 1K or 2K output, stronger prompt adherence, cleaner readable text, and more realistic product, editorial, and marketing visuals.
DeepSeek V4 Flash
DeepSeek V4 Flash is a Flash-tier model for 1M-token long-context text chat, fast non-thinking responses, prompt-guided reasoning, code review, agent planning, and OpenAI-compatible integration.
DeepSeek V4 Pro
DeepSeek V4 Pro is the Pro-tier model for 1M-token long-context text chat, deep reasoning, complex coding, long-horizon agent planning, and OpenAI-compatible integration.
GPT 5.5
OpenAI GPT 5.5 is a premium chat model for long-horizon agentic execution, complex coding, research synthesis, tool-oriented planning, streaming, and OpenAI-compatible production assistants.
GPT Image 2
OpenAI's advanced image generation and editing model for reasoning-guided composition, readable in-image text, photoreal product detail, and controlled reference-based edits.
Xai TTS V1
Xai TTS V1 comes from xAI's Grok voice model family, turning scripts into low-latency, expressive speech with language control, inline delivery tags, five preset voices, and production-ready audio formats.
Claude Opus 4.7
Anthropic's Claude Opus 4.7 is an Opus-tier model for 1M-token long-context reasoning, complex coding, long-horizon agent planning, self-checking workflows, and professional analysis.
Gemini 3.1 Flash TTS
Gemini 3.1 Flash TTS is an expressive text-to-speech model for natural speech, inline Audio Tags, style-guided delivery, 70+ language options, and two-speaker dialogue.
MiniMax Music 2.6
MiniMax Music 2.6 is a MiniMax music generation model for complete prompt-to-song workflows, supporting lyric-assisted vocal tracks, instrumental music, and configurable MP3, WAV, or PCM audio exports.
Happy Horse
Happy Horse is Alibaba's cinematic AI video model for text-to-video, source-image animation, reference-guided scenes, and short video edits, suited to product teasers, social ad concepts, and storyboard previews.
Wan 2.7 Video
Wan 2.7 Video is Alibaba's flexible short-form video API family for text-to-video, start/end image animation, reference-guided generation, and prompt-based video editing with 720p or 1080p output.
Wan 2.7 Image
Alibaba's Wan 2.7 Image is a unified image generation and editing model for prompt-faithful text-to-image output, reference-guided edits, readable visual text, and high-control product, marketing, and design assets.
Tripo3D P1
Tripo3D P1 is Tripo3D's low-poly 3D generation model for creating lightweight, structured asset candidates from prompts or one object image, suited to real-time props, WebGL previews, AR placeholders, and fast mesh iteration.
GPT 5.4
OpenAI GPT 5.4 is a professional chat model for efficient task execution, spreadsheet and document analysis, tool-oriented planning, streaming, and OpenAI-compatible assistants.
Nano Banana 2
Google's fast Nano Banana image model for high-volume generation, reference-guided edits, and readable text output.
Seedream 5.0 Lite
ByteDance Seed's Seedream 5.0 Lite is an efficient image generation and editing model for reasoning-aware prompts, structured layouts, readable text, subject consistency, and reference-guided candidate batches across product, marketing, design, and teaching visuals.
Seedance 2
ByteDance Seed's multimodal video model for native audio-video generation, reference-controlled scenes, multi-shot storytelling, and stable motion in cinematic short-form clips.
Tripo3D H3.1
Tripo3D H3.1 is a high-detail 3D generation family for creating textured assets from prompts, one object image, or 2-4 multiview references, suited to close-up game props, product visualization models, AR previews, and concept asset review.
Kling O3
Kling O3 is Kuaishou Kling AI's flexible video generation family for short clips, combining prompt-only, image-guided, reference-guided, multi-shot, audio, and 4K options for ads, product teasers, storyboard previews, and in-app video creation.
Kling 3.0
Kling 3.0 is Kuaishou Kling AI's video generation family for directed short-form clips, combining prompt or image-guided generation, multi-shot scene structure, Native Audio, and Standard, Pro, or native 4K output choices for ads, product teasers, and storyboard previews.
Kling 3.0 Turbo
Kling 3.0 Turbo is Kuaishou Kling AI's speed-focused 3.0 video variant for prompt, one-image, and multi-shot preview workflows where teams need quick short-form motion tests before heavier production renders.
Kling 3.0 Motion Control
Kling 3.0 Motion Control is Kuaishou Kling AI's reference-video motion transfer model, built to apply body movement, gestures, and performance timing from one video to one character image for controllable character animation, dance transfer, ad creative, and storyboard previews.
Kling O3 Image
Kling O3 Image is a multimodal model for high-quality image generation and editing, supporting text-to-image, image-to-image, reference-guided edits, element control, and 1K, 2K, or 4K output for design, ads, and product visuals.
Claude 4.6
Anthropic's Claude 4.6 API covers Sonnet for balanced coding-agent and production chat work plus Opus for deeper reasoning, long-context analysis, tool use, and computer-use style workflows.
Hunyuan 3D 3.1
From Tencent Hunyuan, Hunyuan 3D 3.1 is a production-grade 3D generation family for turning prompts or object images into cleaner-topology assets, with controllable geometry, PBR material output, and Pro/Rapid options for quality or faster iteration.
Grok Imagine
Grok Imagine is xAI's Aurora-powered visual generation workflow for creating images, editing one reference image, and generating short text-to-video or image-to-video clips with practical mode, ratio, and duration controls.
Meshy 6 3D
Meshy 6 3D is Meshy's 3D generation family for creating textured 3D assets from prompts, one reference image, or 1-4 object views, suited to game props, product AR prototypes, character base meshes, and concept asset iteration.
Seedance 1.5 Pro
Seedance 1.5 Pro is ByteDance Seed's short-form video model for cinematic text-to-video and image-to-video generation, especially when natural motion, first/last frame control, and synchronized audio or lip-sync matter in compact campaign clips.
Wan 2.6
Wan 2.6 is Alibaba's Wan-series video API for text-to-video, single-image animation, and video-to-video restyling, focused on multi-shot 720p or 1080p clips up to 15 seconds for cinematic short-form production.
Runway Gen-4.5
Runway Gen-4.5 is Runway's cinematic short-video generation model for prompt-driven and optional single-image-guided clips where believable motion, physical interaction, and polished visual continuity matter.
GPT 5.2
OpenAI GPT 5.2 is a legacy professional-work model for compatibility reviews, 400K-token long-context synthesis, migration analysis, agentic coding handoffs, and OpenAI-compatible assistants.
Seedream 4.5
ByteDance Seed's Seedream 4.5 is an upgraded image generation and editing model for 2K and 4K production visuals, combining readable text, stronger prompt adherence, subject consistency, and reference-guided control for ecommerce hero shots, ad posters, brand assets, and polished image sets.
Kling 2.6
Kling 2.6 is Kuaishou's Kling AI Video 2.6 model for directed short-form video generation, built for prompt-driven and reference-image-guided clips with stronger camera motion, subject continuity, and realistic scene transitions for product teasers, social ads, and storyboard motion tests.
Kling 2.6 Motion Control
Kling 2.6 Motion Control is Kuaishou Kling AI's reference-video motion transfer model, built to apply body movement, gestures, and performance timing from one video to one character image for controllable character animation, dance transfer, ad creative, and storyboard previews.
Kling O1 Image
Kling O1 Image belongs to Kuaishou's Kling creative model family and focuses on reference-guided image editing for preserving subjects, composition, and style while producing cost-effective product retouching, background changes, material edits, and visual variants.
Z-Image
Alibaba's Z-Image is an efficient text-to-image model for quickly producing realistic visual candidates, helping teams explore product concepts, campaign ideas, and social assets across square, portrait, and landscape aspect ratios.
FLUX.2
FLUX.2 is Black Forest Labs' 32B image generation and editing family for photoreal text-to-image output, multi-reference edits, readable typography, and consistent product, brand, and character visuals.
Nano Banana Pro
Google's premium Nano Banana text-to-image model for cleaner typography, stronger commercial polish, optional web grounding, and higher-end 1K, 2K, and 4K output tiers.
Gemini 3 Series
Google's Gemini 3 Series is a native Gemini chat model family for fast Flash responses, deeper Pro reasoning, coding help, and production assistants that need controllable text generation.
Hailuo 2.3
Hailuo 2.3 is MiniMax's video generation model for text-to-video and start-image-guided clips, focused on fluid motion, expressive characters, style stability, and short cinematic outputs for product concepts, social ads, and storyboard previews.
Veo 3.1 Official
Google Veo 3.1 Official generates cinematic short videos with native audio, prompt, image, first/last-frame, and reference workflows.
Veo 3.1
Google DeepMind's Veo 3.1 video model for native audio, cinematic motion, first/last frame control, reference image guidance, and coherent short-form generation.
Sora 2 Official
OpenAI Sora 2 Official generates short cinematic videos from prompts or one guiding image, with Standard 720p output for iteration and a Pro variant for sharper 720p, 1024p, or 1080p results.
Claude 4.5 Series
Anthropic's Claude 4.5 Series covers Opus for complex coding, Sonnet for balanced production chat, and Haiku for fast high-volume assistant workflows.
Wan 2.5
Wan 2.5 is Alibaba's higher-quality short-video API for text-to-video and single-image animation, suited to polished 5s or 10s clips with optional audio text and output up to 1080p.
Kling 2.5 Turbo Pro
Kling 2.5 Turbo Pro is Kuaishou Kling AI's cost-efficient Pro video model for prompt-only, start-frame-guided, and start/end-frame-guided short clips with controlled cinematic motion for ads, product videos, and storyboard tests.
Seedream 4
ByteDance Seed's Seedream 4 is a unified image generation and editing model for text-to-image, reference-guided edits, and 1K to 4K output, suited to readable posters, product shots, diagrams, educational visuals, and consistent assets.
ElevenLabs Music
ElevenLabs Music is ElevenLabs' text-to-music model for generating full songs, instrumentals, hooks, and soundtrack drafts from natural language, with studio-oriented audio quality and composition-plan control over sections, lyrics, and timing.
Wan 2.2 Fast
Wan 2.2 Fast is Alibaba's rapid short-video API for prompt-to-video and one-or-two-image animation, built for fast 480p or 720p drafts when teams need lower-cost motion iteration before higher-control production work.
Kling Avatar 2.0
Kling Avatar 2.0 is Kuaishou Kling AI's audio-driven talking-avatar model for generating expressive presenter videos from one reference image and one speech audio track.
Hailuo 02
Hailuo 02 is MiniMax's video generation model for physics-aware short clips, supporting prompt-only generation plus optional first-frame or first-and-last-frame guidance for cost-efficient product teasers, social ads, and storyboard motion tests.
Seedance 1.0 Pro
ByteDance Seed's first-generation Seedance Pro video model for prompt-only or single-image-guided 5s and 10s cinematic clips, suited to low-cost 720p and 1080p short-video drafts.
ElevenLabs V3 TTS
ElevenLabs V3 TTS is ElevenLabs' flagship expressive text-to-speech model, built for natural narration, dialogue-style scripts, Audio Tags, 70+ language delivery, and voiceovers that need emotional range instead of plain reading.
Kling 2.1
Kling 2.1 is Kuaishou Kling AI's start-frame-guided image-to-video model, offering Standard and Pro variants for 5 or 10 second clips with stable subject motion, cinematic camera movement, and optional Pro end-frame control for ads, product videos, and storyboard previews.
Flux Kontext
Flux Kontext is Black Forest Labs' in-context image generation and editing family for mask-free text instructions, single-image edits, subject consistency, style transfer, and text-aware visual updates.
Flux Dev
Flux Dev is Black Forest Labs' open-weight FLUX.1 Dev image model for high-quality text-to-image generation, strong prompt adherence, realistic detail, and single-reference image-to-image guidance.
Flux Schnell
Flux Schnell is Black Forest Labs' open-weight FLUX.1 Schnell image model for fast, low-cost text-to-image drafts, prompt exploration, and batch creative iteration.
ElevenLabs TTS Turbo 2.5
ElevenLabs TTS Turbo 2.5 is ElevenLabs' Turbo v2.5 low-latency TTS model for natural speech across 32 languages, 40000-character scripts, and high-volume product prompts, support flows, localization drafts, and narration where speed and cost matter most.
AI Music
AI Music is a Suno-style prompt-to-song music generation model for turning prompts, style guidance, titles, and instrumental or vocal settings into full songs, background tracks, songwriting drafts, and music product prototypes.