Inception claims Mercury 3 beats GPT-6 Luna and Claude Haiku 5.5
Original title: Inception's Mercury 3 diffusion model claims to beat GPT-6 Luna and Claude Haiku 5.5 at a fraction of the wait
Inception's own tests give Mercury 3 eight of 12 wins and 1,068 tokens per second, with no independent verification yet.
Inception announced Mercury 3, a diffusion language model it calls the most capable of its kind released so far. The company says it beats GPT-6 Luna and Claude Haiku 5.5 on a range of benchmarks while generating many tokens at once, but its own table is mixed and a shorter overview says quality is on par or better. Mercury 3 is pitched for agents, search, support and tutoring, with launch pricing cut in half and availability on Inception, OpenRouter, Baseten and Vercel.
Unconfirmed: Speed, benchmark and partner figures come only from Inception; the parameter count is undisclosed and no independent tests exist.
- Mercury 3 scored 76.84 on IFBench, 15 points above GPT-6 Luna.
- On SWE-Bench Atlas it scored 51.08, versus 33.33 and 32.26 for rivals.
- Claude Haiku 5.5 won DeepSWE and long-text retrieval tests.
- Listed speed is 1,068 tokens per second, versus about 240 and 130.
- Regular prices are $0.20 per million input tokens and $0.70 output.
- Vendors
- Inception
- Organizations
- OpenRouterBasetenVercel
- AI models
- Mercury 3GPT-6 LunaClaude Haiku 5.5
From the original article
3 min read at the-decoder.com
Inception says Mercury 3 is the most capable diffusion language model released so far. The company claims it runs four to eight times faster than GPT-6 Luna and Claude Haiku 5. 5 and beats both in eight of 12 benchmarks.
By Manuel Uth · the-decoder.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.