Dev Tools Universe

A dedicated AI editor with its own Composer and Grok models built in.

Made by Anysphere, a SpaceX subsidiary since August 2026. Plans include Cursor's own Composer 2.5 and Grok models; Claude, GPT and Gemini are available on Premium ($30/mo) and up, billed at API list price. Runs agents in parallel, locally or in the cloud.

The former Windsurf editor, now Cognition's Devin Desktop.

Cognition acquired Windsurf in July 2025 and renamed it Devin Desktop in June 2026. The Devin Local agent replaced Cascade, VS Code extensions and keybindings still work, and the same plans cover Devin's cloud agent. The free tier has a light quota.

Anthropic's agentic coding tool — terminal-first and extensible.

Anthropic's coding agent, running on Claude Opus 5.5 by default on paid plans and the API. It reads files, runs commands and checks its own work, and extends with skills, subagents, hooks, MCP servers and plugins. Included on every Claude plan except Free. (This hub was built with it.)

AI in the IDE you already use, with SSO & audit trails.

GitHub's (Microsoft) assistant: completions, chat and agent mode, plus the Copilot cloud agent, CLI and app on every plan. Models come from OpenAI, Anthropic and Google. Free and Student tiers; paid plans from Pro at $10/mo, with usage metered in GitHub AI Credits since June 2026.

Google's agent-first development platform — generally available.

Google's agent-first IDE, CLI and SDK. Version 2.0 added subagents and custom agents. Gemini 3.8 Flash runs on every plan, and Claude Sonnet 5.5 and Claude Opus 5.5 on paid tiers. The individual plan is free; Google AI Pro or Ultra raise the limits.

A performance-focused editor with AI built in.

A lean editor with built-in AI and agent features. The Personal plan is free; Pro is $10/mo and includes $5 of model tokens. A good pick when responsiveness matters more than a huge extension catalogue.

Open-source AI inside the universal editor.

Stay in plain VS Code and add Cline, the open-source (Apache-2.0) coding agent with 5M+ installs. Free for individuals — bring your own model key and pay only for inference, with full control over your stack.

Anthropic's recommended starting model — strong coding and agentic work, 1M-token context.

Anthropic's docs recommend it as the starting model for most workloads. Anthropic says it performs at the level of Claude Fable 5.1 on most work and costs about 40% less to run than Claude Opus 5. $4 in / $20 out per MTok; cache reads $0.20; fast mode $8 / $40. The default model in Claude Code.

OpenAI's flagship for complex reasoning, coding and computer use.

OpenAI's flagship for complex reasoning, coding, agentic work and computer use. For lower cost, GPT-6.1 Sol is the balanced tier and GPT-6 Luna the cheapest.

Google's flagship reasoning model, 1M-token context.

Google's flagship Gemini model, generally available since 13 October 2026, with a 1M-token context and long-decode output. Paid tier only; $2 in / $10 out per MTok through 30 Nov 2026, then $4 / $20. For cheap, high-volume work, Gemini 3.8 Flash has a free tier.

Meta's current open-weight model — small enough for one GPU.

Meta's open weights now come from the Muse family. Muse Glimmer 30B (Apache 2.0, August 2026) is distilled from the closed Muse Spark and runs on a single consumer GPU. Llama 4 (Scout, Maverick) (April 2025) was Meta's last Llama release — still downloadable, with a 10M-token context on Scout, but no longer updated.

Z.ai's coding-focused open-weight flagship.

Z.ai calls it "the most capable open-weights model for coding". Free to use under its own licence, not an OSI open-source one: Custom GLM-5.3 licence (MIT-style, plus a security review for model-as-a-service firms over $10B revenue); GLM-5.3-Flash is plain MIT.

Very low API prices, MIT-licensed weights.

DeepSeek's fast multimodal model with a 1M-token context. MIT weights; API $0.15–0.30 in / $0.60–1.20 out per MTok (off-peak / peak). The flagship DeepSeek-V4-Pro (text only) is also MIT. Self-hostable for budget-conscious, high-volume use.

Dev favorite. Cheap, honest pricing, good UI.

Popular among developers. Transparent pricing, no dark patterns on renewal. Free WhoisGuard privacy. Solid DNS management. Great for personal projects and startups.

Most recognized name. But notorious for upselling.

Massive company, owns millions of domains. Good for non-techies. But known for aggressive upsells, confusing UI, and charging extra for privacy. Avoid if you know what you're doing.

At-cost pricing. Best if you already use Cloudflare.

Cloudflare sells domains at wholesale cost — no markup. Free privacy. Best DNS in the world. But you can mainly transfer existing domains rather than register every new TLD. Ideal for serious devs.

Formerly Google Domains. Now run by Squarespace.

Google Domains was beloved for simplicity, then sold to Squarespace in 2023. Clean UI, fair pricing. If you were on Google Domains, you're now on Squarespace Domains.

Quirky, cheap, great for .io and .ai domains.

A developer darling. Fun branding, genuinely cheap prices (especially for .io, .ai, .dev). Free WHOIS privacy. Good DNS management. Great for indie hackers and side projects.

Clean, no-upsell experience. Simple and honest.

A clean, no-nonsense registrar. No upsells, no confusion. Slightly pricier but great customer support. Good for people who hate the GoDaddy checkout maze.