Inception's New AI Writes 8x Faster, Making Chatbots Feel Instant
This could make AI assistants feel as fast as a human—no more waiting.
Inception introduced Mercury 3 on Oct. 19, claiming faster text generation and competitive benchmark performance against models from OpenAI and Anthropic.
According to Inception, Mercury 3 is a diffusion LLM that reaches 1,068 tokens per second on Nvidia GPUs and matches or outperforms OpenAI’s GPT-6 Luna and Anthropic’s Claude Haiku 5.5 across a range of benchmarks. It supports a 1 million-token context window, up from 260,000 in Mercury 2.5.
Standard pricing is $0.20 per million input tokens and $0.70 per million output tokens, with introductory prices of $0.10 and $0.35; the article says Inception’s launch announcement does not specify when the introductory discount ends. Inception also published early-access partner reports on browser navigation, education hints and smart-glasses search, but the article notes those reports illustrate potential uses rather than independently establish a general performance advantage.
- Mercury 3 generates text up to 8 times faster than rivals, reaching over 1,000 words per second.
- It's cheaper, at $0.20 per million input words and $0.70 per million output words, with introductory discounts.
- Faster AI could make chatbots, tutors, and voice assistants feel more instant and natural.
Why It Matters
Faster AI means less waiting and more natural conversations with chatbots, tutors, and virtual assistants.