Inception has introduced Mercury 3, which it calls the most capable diffusion-based LLM released to date. Across a range of benchmarks, Mercury 3 outperforms GPT-6 Luna and Claude Haiku 5.5 while running 4–8x faster at 1,068 tokens per second on NVIDIA GPUs. The model is more than 40x larger than Mercury 1 and is vastly more intelligent, yet just as fast. Its context window has grown to 1M tokens, up from 260K for Mercury 2.5.
Pricing is $0.20 per million input tokens and $0.70 per million output tokens. At launch, Mercury 3 is 50% off, at $0.10 per 1M input and $0.35 per 1M output. Cached tokens are 90% off the input price. The company says the scaling laws that drove progress in autoregressive models also hold for diffusion, and that intelligence across standard benchmarks has risen 3x from Mercury 1 (mini) to Mercury 3.
Early-access partners are already using Mercury 3 for latency-sensitive work. In one test, a browser-use agent built on Mercury 3 completed 50% of 50 live Mind2Web navigation tasks, significantly more than GPT-6.1-Sol, at a fraction of the per-token cost. Other partners report near-real-time performance in interactive education, where hints and feedback now return in about a second, and in real-time search, where Octen measured a median generation time of 0.9 seconds and a median end-to-end response of 2.2 seconds.
Support teams are also adopting the model. Lorikeet uses Mercury to keep conversational customer responses accurate for regulated businesses, where a wrong answer is costly.
Source: Hacker News · Summarized by HeadlinesBriefing