Mercury Edit 2
Inception's diffusion model for code autocomplete and next-edit prediction
Editorial take
Mercury Edit 2's standout capability is fast next-edit and autocomplete predictions from a diffusion model. It is best suited for teams building editors, ide plugins or code tools that need fast autocomplete and next-edit suggestions. A free plan is available, with paid plans starting at $0.25/1M input tokens.
What is Mercury Edit 2?
Mercury Edit 2 is a diffusion language model (dLLM) from Inception built for the most latency-sensitive parts of coding: inline fill-in-the-middle autocomplete and next-edit prediction. Instead of writing one token at a time, it generates tokens in parallel; Inception says its diffusion models run at 1,000+ tokens per second. For next-edit suggestions it reads your recent edits, the current file and recently viewed code, then proposes the change you are likely to make next. Inception says training on human accept and reject feedback made its edits accepted 48% more often than its previous next-edit model and 27% more selective about which edits it shows.
Released on March 30, 2026, it is available through the Inception API Platform on dedicated FIM and edit endpoints, with official Python and TypeScript SDKs, and as an edit prediction provider in the Zed editor and the ProxyAI plugin. Every new Inception account gets a one-time credit of 100 million free tokens shared across models. Inception's flagship general model is now Mercury 3 (October 2026); Mercury Edit 2 remains its code editing model.
Best for
Teams building editors, IDE plugins or code tools that need fast autocomplete and next-edit suggestions
Key strength
Fast next-edit and autocomplete predictions from a diffusion model
Score breakdown (out of 5)
Pros & Cons
π Pros
- βBuilt specifically for low-latency autocomplete and next-edit suggestions
- βLow per-token price ($0.25 input and $0.75 output per 1M tokens)
- β100 million free tokens for new accounts
- βWorks in Zed with just an API key
π Cons
- βDeveloper API only - no consumer product
- β32K context window is far smaller than general-purpose LLMs
- βText only, with no tool calling or structured outputs
- βNext-edit requests need a specific tagged prompt format
Key Features
- β Diffusion-based architecture (not autoregressive)
- β Fill-in-the-middle (FIM) autocomplete endpoint
- β Next-edit prediction from recent edits, file context and viewed snippets
- β 32K context window, 8,192 max output tokens
- β Official Python and TypeScript SDKs
- β Edit prediction provider in Zed and ProxyAI
- β Inception reports 1,000+ tokens per second for its diffusion models
Mercury Edit 2 Pricing
β Mercury Edit 2 has a free plan - no credit card required to start.
Free credit
- β100 million free tokens for new accounts, shared across Inception models
- βFree tier rate limits
API
- β$0.25 per 1M input tokens ($0.025 cached)
- β$0.75 per 1M output tokens
- β32K context, 8,192 max output
Mercury Edit 2 vs Competitors
Developer resources
Related Tools
Claude is the best AI assistant for long-document work and research where accuracy matters. Its context window of up to 1M tokens and resistance to sycophancy make it the honest choice when you need a thinking partner rather than an agreeable assistant. Start on the free tier; the Pro plan at $20/month is worth it if you regularly hit the free usage limits, and Max ($100/month) or Ultra ($200/month) suit heavier daily use.
ChatGPT is the most capable all-in-one AI assistant available. Its combination of image generation, voice mode, code execution, web browsing, and a large custom GPT ecosystem makes it the default choice for users who want one tool that handles everything. The tradeoff is that it is rarely the deepest tool on any single task.
Cursor is the most powerful AI code editor available. Its multi-file editing is ahead of plugin-based alternatives, making large-scale refactors practical in ways GitHub Copilot cannot match. At $30/month for Premium (Grok models included, other models billed at API rates) it is a meaningful upgrade for any developer doing more than small isolated changes.
GitHub Copilot is the right starting point for developers adopting AI-assisted coding. It integrates into whatever editor you already use, requires no workflow change, and has a usable free tier. The Pro plan at $10/month is the lowest-cost path to unlimited AI completions, though chat and agent work now draw from a monthly credit allowance; heavy agent users should look at Pro+ ($39) or Max ($100).
This page contains affiliate links. We may earn a commission at no extra cost to you. Learn more.