Features Use Cases Models Pricing Download Sign In Get Started

Every model, one workspace

Pick the right model for each assistant and each task — from the fastest and most economical to the most capable. You can change an assistant's model at any time.

Language models

30+ chat and reasoning models across the four leading providers. Below are the current highlights — earlier versions remain available in-app.

Model Provider Best for
Claude Opus 4.8 Anthropic Deep reasoning and complex analysis
Claude Sonnet 4.6 Anthropic Balanced quality and speed for everyday work
Claude Haiku 4.5 Anthropic Fast, lightweight responses
GPT-5.5 OpenAI Top-tier reasoning and tool use
GPT-5.4 OpenAI Versatile default for most tasks
GPT-5.4 Pro OpenAI The hardest problems, maximum quality
GPT-5.4 Mini OpenAI Fast and economical
GPT-5.4 Nano OpenAI Ultra-fast for simple tasks
Gemini 3.1 Pro Google Long-context, multimodal reasoning
Gemini 3.5 Flash Google Fast multimodal at scale
Gemini 3 Flash Google Quick everyday tasks
Grok 4 xAI Real-time knowledge and reasoning
Grok 4.3 xAI High quality with a large context window
Grok 4.1 Fast xAI Very fast, with a huge context window

Image generation

Generate and edit images in many styles and resolutions, directly in your conversations.

GPT Image

OpenAI

Imagen

Google

Gemini Image

Google

Grok Imagine

xAI

Flux

Black Forest Labs

Ideogram

Great for text in images

HiDream

Open image model

Stable Diffusion

Open image model

Video generation

Create video clips and refine them with built-in editing, avatar, and lip-sync tools.

Veo

Google

Sora

OpenAI

Kling

Text- and image-to-video

Grok Video

xAI

Wan

Reference-to-video

Seedance

Cinematic motion

HeyGen Avatars

Talking-avatar video

Lip-sync & editing

Trim, subtitle, upscale

Voice & audio

Turn text into natural speech and transcribe audio accurately.

Text-to-speech

Lifelike voiceovers from OpenAI and ElevenLabs, with a wide range of voices and optional word-level timing.

Speech-to-text

Accurate transcription with OpenAI Whisper and ElevenLabs Scribe — used for voice notes, audio files, and meeting content.

Coming soon

We're continually adding models. Open-source language models are on the way.

Llama 4

Meta

DeepSeek R1

Open reasoning model

Qwen

Alibaba

All model usage consumes AI credits based on the real token and compute cost of each request — lighter models and lower-resolution media cost fewer credits than premium models or high-quality video. Team messaging between people is always free. See pricing for plans and example costs.

Try them all in one place

Sign up and give each assistant the model that fits the job.

Get Started