AI models
behind one API key
33 image, video and language models. Test each one in a live playground, then move the same request into production with pay-as-you-go API billing.
ByteDanceJust added
Wan 3.0
Text · Image · References · Audio
Alibaba's newest video model: up to 30 seconds of 1080p video in one take, with dialogue, ambience and sound effects.
Try model
Krea 3
Text to image · Edit
Krea's newest model — aesthetic generation with built-in prompt expansion, and edits that change only what you ask.
Try model
Nano Banana 2.1
Text to image · Edit · 4K
Google's newest Nano Banana — 4K output, extreme aspect ratios, and web-grounded generation at Nano Banana 2 pricing.
Try model
GPT Image 2.5
Text to image · Edit
OpenAI's newest image model — five quality tiers, and edits that leave the rest of the frame alone.
Try modelAll models
Krea 3
Text to image · Edit
Krea's newest model — aesthetic generation with built-in prompt expansion, and edits that change only what you ask.
Try modelNano Banana 2.1
Text to image · Edit · 4K
Google's newest Nano Banana — 4K output, extreme aspect ratios, and web-grounded generation at Nano Banana 2 pricing.
Try modelGPT Image 2.5
Text to image · Edit
OpenAI's newest image model — five quality tiers, and edits that leave the rest of the frame alone.
Try model
Midjourney V8
Text to image · 4 per run
Midjourney V8.2 without Discord: four variations per run, native 2K HD, style references, and stylize, chaos and weird controls.
Try model
Nano Banana Pro
Text to image · Edit · 4K
High-fidelity generation and editing with up to 4K output.
Try model
Nano Banana 2
Text to image · Edit
Fast, versatile image generation for everyday creative workflows.
Try model
GPT Image 2
Text to image · Edit
Precise generation and editing with three selectable quality tiers.
Try modelSeedream V5 Pro Unrestricted
Text to image · 10 references
Create 1K or 2K visuals with strong typography and up to ten reference images.
Try modelSeedream 5
Text to image · Edit
High-fidelity generation and reference-guided editing in Pro and Lite variants.
Try model
Ideogram Remove Background
Image editing
Remove backgrounds while preserving hair, glass, fine text, and difficult edges in a transparent PNG.
Try modelWan 3.0
Text · Image · References · Audio
Alibaba's newest video model: up to 30 seconds of 1080p video in one take, with dialogue, ambience and sound effects.
Try model
MiniMax H3 Max
Text · Image · Frame control
Polished 5–15 second video generation through one JSON endpoint, or test it in the live generator.
Try model
Midjourney V8 Video
Image to video · 4 clips per run
Animate a still image into four 5-second clips in one run, with low or high motion. No Discord needed.
Try modelSeedance 2.5
Text · Image · Omni references · Audio
Cinematic 480p or 720p video up to 30 seconds from text, images, first and last frames, or omni image, video and audio references.
Try modelSeedance 2.5 Unrestricted
Video · Audio · Frame control
4–30 second clips from text or images with fewer creative restrictions, first and last frames, and optional synced audio.
Try model
Pixverse V6
Video · Styles · Audio
Video from text or a still image, steered with creative style presets, and clip extension when you need longer motion.
Try model
Kling 4.0 & Flash
Preview · 30s multi-keyframe
Compare Kling 4.0's 30-second, multi-keyframe features with the faster Flash variant. Not available on imageat yet.
See preview
Qwen 3.8 27B Uncensored
262K context · OpenAI compatible
A 27B long-context model with optional thinking mode, an OpenAI-compatible chat endpoint, and a browser playground.
Open playgroundGemini 4 Argon
Multimodal reasoning for long, multi-step work
Google's Gemini 4 Argon reasons over text, images, video, and audio for long, multi-step coding, research, and document analysis, with optional Google Search and Maps grounding.
Open playgroundGPT-6 Astra
Frontier reasoning for agents, code, and creative production
OpenAI's GPT-6 Astra combines long-context reasoning, optional image understanding, web search, and tool calling for coding, research, and document workflows.
Open playground
Qwen 3.8 Flash Next GGUF
Fast long-context chat in an optimized GGUF build
Unsloth's optimized Qwen 3.8 Flash Next GGUF deployment for long-context chat, coding, research, and optional thinking-mode responses.
Open playgroundQwen 3.8 27B Obliterated
Reduced-refusal long-context reasoning and coding
A 27B Qwen 3.8 variant with reduced refusal behavior for long-context chat, coding, research, and controllable thinking traces.
Open playgroundQwen 3.8 27B
Long-context reasoning for agents, research, and code
The standard Qwen 3.8 27B deployment for long-context reasoning, agentic coding, research, and controlled chat generation.
Open playgroundGemini 3.8 Flash
Multimodal reasoning for long-horizon agents and coding
Google's Gemini 3.8 Flash reasons over text, images, audio, video, and PDFs, then returns strong text answers for long-horizon agent and coding workflows.
Open playgroundGemini 3.7 Flash
Fast multimodal reasoning for coding and agents
Analyze text, images, audio, and video with adjustable thinking depth and long-form text output for coding, agent, and knowledge workflows.
Open playgroundGPT 5.6 Sol
Flagship reasoning for demanding work
OpenAI's premium GPT 5.6 tier for complex coding, research, visual analysis, and detailed multi-step answers.
Open playgroundGPT 5.6 Terra
Balanced multimodal intelligence for everyday work
A balanced GPT 5.6 model for analysis, writing, coding, and grounded answers from text and images.
Open playgroundGPT 5.6 Luna
Speed-first chat and vision Q&A
A fast GPT 5.6 tier for high-volume chat, screenshot analysis, image captioning, and concise visual answers.
Open playground
Grok 4.5
Long-context reasoning with tool use
xAI's flagship reasoning model for coding, long-context analysis, tool calling, and image understanding.
Open playground
GLM 4.7 Flash
Fast, efficient general-purpose language model
A speed-focused GLM model for chat, extraction, structured writing, coding assistance, and high-volume language tasks.
Open playgroundGPT-OSS 20B
Compact open-weight reasoning
A compact 20B open-weight model suited to efficient reasoning, coding, structured generation, and controllable deployments.
Open playground
Claude Sonnet 5
Balanced frontier intelligence for agents and coding
Anthropic's balanced Claude 5 tier for software work, document analysis, agent workflows, and nuanced long-form writing.
Open playgroundClaude Opus 5
Maximum-capability Claude reasoning
Anthropic's highest-capability Claude 5 tier for difficult reasoning, complex coding, research, and long-horizon agent tasks.
Open playgroundOne balance, every model
Compare per-model prices, see how API billing works, and create a key to call any model above from your own app. More text, image, video and audio endpoints are added as they launch.