Inception Mercury 3

Name: Inception Mercury 3
Author: Inception

Mercury 3 is Inception's newest diffusion-based language model, with a 1M-token context window, an OpenAI-compatible API, and a list price of $0.20 per million input and $0.70 per million output tokens.

Visit Inception Mercury 3 ↗

Paid (usage-based)ProprietaryREST API (OpenAI-compatible)by Inception

Why it matters

For teams building latency-sensitive apps such as support chat, voice agents, tutoring or browser agents, it offers a fast model that drops into existing OpenAI-style code, at a low per-token price.

Highlights

Inception reports about 1,068 tokens per second on NVIDIA GPUs and says Mercury 3 matches or beats GPT-6 Luna and Claude Haiku 5.5 on its benchmarks while running 4-8x faster.unverified
Context window grows to 1M tokens, up from 260K in Mercury 2.5.
Inception says the model is more than 40x larger than Mercury 1, presenting it as evidence that text diffusion models scale.
OpenAI API compatible and available on the Inception API Platform, OpenRouter, Baseten and Vercel; new API users get 100 million free tokens.
Launch pricing is 50% off ($0.10 input / $0.35 output per 1M tokens), and cached input tokens are 90% off.

Taken from the project's own README, docs or launch page. Claims about speed or accuracy are the maker's unless stated otherwise.

Try it if: Your chatbot or agent feels slow and you want to test a faster model without rewriting your OpenAI-style code.

Worth knowing: Speed and quality comparisons, including the Mind2Web result, come from Inception and its early-access partners. The launch discount's end date isn't stated.

Quick facts

Pricing	$0.20 per 1M input and $0.70 per 1M output tokens ($0.10 / $0.35 during the launch discount); cached input 90% off; 100M free API tokens to startnot verified
Free or paid	Paid (usage-based)
License	Proprietary
Language	REST API (OpenAI-compatible)
Platform	Inception API Platform, OpenRouter, Baseten, Vercel; Mercury Chat in the browser
GitHub stars	n/a (not a GitHub project)
Maker	Inception

Related ranking

See how established tools compare in Best AI Assistants for Everyday Work. New releases aren't scored until we review them against the rubric.

What we couldn't verify

highlights[0] (speed and benchmark claims are Inception's)
quick_facts.pricing (launch discount end date)

Checked 2026-10-22 from the model's own pages. Prices and star counts change, so confirm on the official site. Listing here is not a ranking or an endorsement.