🤖Mercury 3: 40x Larger, 4-8x Faster LLM
Mercury 3 is 40x bigger and 4-8x faster than its predecessor
TL;DR
Mercury 3, the latest diffusion-based LLM, is 40x larger and 4-8x faster than Mercury 1, setting new benchmarks in speed and intelligence. It's now available at a launch discount, making it a game-changer for real-time applications.
Mercury 3, the most capable diffusion-based LLM to date, outperforms GPT-6 Luna and Claude Haiku 5.5 across various benchmarks, running 4-8x faster at 1,068 tokens per second. This model is not just about speed; it's also 40x larger than its predecessor, offering unparalleled intelligence and performance. Developers and teams using real-time applications like voice support, tutoring, and web navigation will benefit from Mercury 3's instant feedback and reasoning capabilities. At launch, Mercury 3 is 50% off, priced at $0.10 per 1M input and $0.35 per 1M output, making it accessible for a wide range of use cases.
Key Points
Mercury 3 is 40x larger than Mercury 1, with 1,068 tokens per second on NVIDIA GPUs.
Runs 4-8x faster than GPT-6 Luna and Claude Haiku 5.5 at 1,000+ tokens per second.
Priced at $0.10 per 1M input and $0.35 per 1M output, 50% off at launch.
Cached tokens are 90% off the input price, reducing costs for frequent users.
Available on Inception API Platform and through OpenRouter, Baseten, and Vercel.
Why It Matters
If you're running real-time applications like voice support or interactive education, Mercury 3's 40x increase in size and 4-8x speed boost over Mercury 1 means faster, more intelligent responses. At $0.10 per 1M input, it's now more affordable, making it a compelling choice for teams looking to enhance their AI capabilities.
Comments
Be the first to comment
Enjoyed this article?
Get it daily. 7am. Free. Reads in 5 minutes.
Join 3,566 builders reading daily.