For $20/month, you can access cutting edge AI tools and features with a full mobile app to go with it.

Backed by superb performance on Livebench, ARCAGI, and Fiction.Livebench benchmarks, GPT-5 is the best model for general use.

As frontier models get better, this category is becoming more of a subjective call. However, based on extensive testing, Anthropic's Claude 4 Sonnet consistently produces high-quality written content that is coherent, engaging, and contextually relevant. Its ability to understand nuanced prompts and generate human-like text makes it an excellent choice for writing tasks.

The free access tier to Google's Gemini 2.5 Pro provides users with a powerful AI model for general purposes, though it comes with certain restrictions. Notably, the free tier has a smaller context window (32,000 tokens compared to paid tiers) and is subject to daily query limits, which will restrict general usage and processing of very large documents. That being said, Gemini 2.5 Pro is a solid option for users who cannot opt for a premium plan.

Nano Banana is great at making targeted edits to images while preserving all other features in an image. For example, you can change a subject's shirt color while keeping all facial features (and everything else) exactly the same.

Gemini Veo3 is the first widely available AI model that can generate full video clips with audio from simple text prompts. Gemini Veo3 produces impressive results for short clips and is easy to use.

OpenAI's GPT-5-Codex is the best model for writing and editing code. It is available at https://chatgpt.com/codex, VS Code Extension (so Cursor and Windsurf too), and via the Codex CLI. (`npm install -g @openai/codex`)

Lovable is a no-code platform that makes it easy for anyone to build out an app with both backend and frontend frameworks all in one place. Especially welcoming to non-technical users.