Inception Labs released Mercury 2.5, a diffusion-based large language model, on September 8, achieving 1,107 tokens per second on commodity NVIDIA GPUs at $0.04 per million input tokens. Unlike autore…
Inception Labs launched Mercury 2.5 on September 8th, a diffusion-based language model API that the company claims generates 1,107 tokens per second on widely available Nvidia GPUs, with a 260,000-tok…
Latency, not intelligence, is now the primary bottleneck for interactive AI applications, according to a blog post by smiz that argues models must respond within 100ms to enable real-time interfaces. …
A developer explains how diffusion language models, which generate text by filling in masked tokens rather than predicting one token at a time, are emerging as a faster alternative to autoregressive m…
Diffusion language models, which generate entire sequences at once and iteratively refine them, have become competitive with autoregressive models on quality since 2024, with releases from Inception L…
Celeris, a new AI research lab backed by Lightspeed Venture Partners, has released Celeris-1, a language model built on diffusion architecture that claims to generate output 24 times faster than GPT-5…
NVIDIA released the 30B-parameter Nemotron-Labs TwoTower language model on July 1 that achieves 2.42x faster inference than its autoregressive baseline without retraining from scratch, by splitting a …
Inception Labs' Mercury 2 diffusion language model outperforms Google DeepMind's DiffusionGemma on reasoning benchmarks while achieving parallel text generation speeds of over 1,000 tokens per second.…