Inception
1 model built · 1 served
Inception AI is a Palo Alto startup founded in 2024, built around a different bet on how language models should generate text. Instead of the token-by-token autoregressive approach every other major lab uses, Inception built Mercury, the first commercial-scale diffusion-based large language model — generating text more like an image diffusion model refines pixels, in parallel rather than sequentially. On OpenKey, Inception has one model: Mercury 2, released March 2026. It's a narrow lineup by design; this is a research-driven lab still proving out an architecture, not a full-catalog provider.
Lineup
There's no lineup to navigate — Mercury 2 is the only Inception model on OpenKey, so it's the flagship, the budget option, and the newest release all at once. It runs $0.25 per 1M input tokens and $0.75 per 1M output tokens, with a 128K context window, handling text-to-text tasks. On OpenKey that's $0.2575 in / $0.7725 out per 1M tokens after the flat 3% fee ($0.25 × 1.03, $0.75 × 1.03). If you want to test diffusion-based generation instead of the usual autoregressive models, Mercury 2 is the one to try — there's no cheaper or free alternative from this lab yet.
- Mercury 2Mar 4, 2026 128K $0.258 $0.772 ctx 128K · in $0.258 · out $0.772
Prices per 1M tokens, flat 3% fee included.
Models Inception serves
Beyond its own models, Inception hosts 1 model in the OpenKey catalog. Prices are Inception’s own list price per 1M input tokens, before our fee.
From the OpenRouter catalog, Jul 28, 2026. All providers
Privacy policy ↗Terms of service ↗
Questions
- What does Inception's pricing look like?
- Just one model, one price point: Mercury 2 at $0.25 per 1M input tokens and $0.75 per 1M output tokens from the provider. On OpenKey that's $0.2575 and $0.7725 per 1M tokens respectively, after the flat 3% fee ($0.25 × 1.03 and $0.75 × 1.03). There's no free tier and no cheaper sibling model to fall back on.
- Which Inception model should I use?
- Mercury 2 — it's the only one available. It's a text-to-text model with a 128K context window, released March 2026. Since Inception doesn't publish a broader lineup on OpenKey, there's no flagship-vs-budget decision to make; Mercury 2 is the default and only option.
- Can I access Mercury 2 through OpenKey's API?
- Yes — Mercury 2 is available through OpenKey's single API key alongside the other models across OpenKey's 52 labs. You get provider list pricing plus the flat 3% fee, with no separate account or contract needed with Inception AI directly.