Get started
Models
Every model one key can call, and the endpoints that serve it.
Name the model in each request. Prices are on the console’s models page, the source of truth for what each model costs.
Decisions
| Model id | Endpoints | Notes |
|---|---|---|
tod-picker | POST /v1/systemone, POST /v1/pick | Typed decisions with a calibrated probability per option. Text and up to 8 images per state. Input tokens only. |
Text generation
| Model id | Endpoint | Notes |
|---|---|---|
claude-fable-5-1 | POST /v1/chat/completions | Anthropic’s most capable model, for demanding reasoning and long-horizon agentic work. 1M context. |
claude-opus-5-5 | POST /v1/chat/completions | Long-running agentic coding and knowledge work. 1M context. |
claude-sonnet-5-5 | POST /v1/chat/completions | The best combination of speed and intelligence in the Claude line. 1M context. |
gemini-4-argon | POST /v1beta/models/gemini-4-argon:generateContent | Google’s flagship reasoning model. 1M context. |
gemini-3.8-flash | POST /v1beta/models/gemini-3.8-flash:generateContent | The most capable Flash model: fast and inexpensive. |
gemini-3.5-flash-lite | POST /v1beta/models/gemini-3.5-flash-lite:generateContent | The fastest, cheapest Gemini model, for high-volume work. |
Image generation
| Model id | Endpoint | Notes |
|---|---|---|
nano-banana-2 | POST /v1beta/models/nano-banana-2:generateContent | Gemini 3.1 Flash Image: generation and editing from text and images. |
nano-banana-2.1 | POST /v1beta/models/nano-banana-2.1:generateContent | Gemini Nano Banana 2.1: generation and editing from text and images. |
nano-banana-pro | POST /v1beta/models/nano-banana-pro:generateContent | Gemini 3 Pro Image: higher-fidelity generation and editing, up to 4K. |
nano-banana-2-lite | POST /v1beta/models/nano-banana-2-lite:generateContent | Gemini 3.1 Flash-Lite Image: faster, cheaper generation and editing at 1K. |
grok-imagine-image | POST /v1/images/generations, POST /v1/images/edits | xAI’s Grok Imagine image model: text-to-image and editing. |
grok-imagine-image-2 | POST /v1/images/generations, POST /v1/images/edits | Grok Imagine Image 2.0, with a quality setting. |
grok-imagine-image-quality | POST /v1/images/generations, POST /v1/images/edits | Grok Imagine’s quality-focused image model. |
See Nano Banana and Grok Imagine.
Video generation
| Model id | Endpoints | Notes |
|---|---|---|
grok-imagine-1 | POST /v1/videos/generations, GET /v1/videos/{request_id} | xAI’s Grok Imagine: text-to-video and image-to-video, 1–15 s with audio, up to 720p. |
grok-imagine-1.5 | POST /v1/videos/generations, GET /v1/videos/{request_id} | xAI’s Grok Imagine 1.5: the same at native 1080p, plus reference images, keyframes and a fixed last frame. |
See Grok Imagine.