← all releases
Inception Labs·Feb 27, 2026·v2.0·paid

Mercury 2

Mercury 2 is a cutting-edge diffusion-based large language model (dLLM) developed by Inception Labs, designed to deliver ultra-fast and efficient AI-driven text generation. By leveraging diffusion technology, Mercury 2 generates multiple tokens simultaneously, achieving speeds over 1,000 tokens per second on NVIDIA Blackwell GPUs, significantly outperforming traditional autoregressive models.

LLMs
visit website
overview

Mercury 2 is a cutting-edge diffusion-based large language model (dLLM) developed by Inception Labs, designed to deliver ultra-fast and efficient AI-driven text generation. By leveraging diffusion technology, Mercury 2 generates multiple tokens simultaneously, achieving speeds over 1,000 tokens per second on NVIDIA Blackwell GPUs, significantly outperforming traditional autoregressive models.

key features
  • 01Diffusion-based architecture for parallel token generation
  • 02Over 1,000 tokens per second processing speed on NVIDIA Blackwell GPUs
  • 03Competitive performance with leading speed-optimized models
  • 04Tunable reasoning capabilities
  • 05128K context window support
  • 06Native tool usage
  • 07Schema-aligned JSON output
related productsbrowse all →
Protenixopen sourceProtenix is an open-source, trainable PyTorch implementation of AlphaFold 3, designed for high-accuracy biomolecular structure prediction. It aims to advance accessible and extensible research tools for the computational biology community. ([github.com](https://github.com/bytedance/Protenix?utm_source=openai))Step-Audio-R1open sourceStep-Audio-R1 is an advanced audio language model developed by StepFun AI, designed to enhance audio reasoning capabilities by grounding its reasoning in acoustic features. It introduces Modality-Grounded Reasoning Distillation (MGRD), an iterative training framework that shifts the model's reasoning from textual abstractions to acoustic properties, effectively addressing the 'inverted scaling' problem where performance degrades with longer reasoning. This model has demonstrated superior performance across various audio understanding and reasoning benchmarks, surpassing models like Gemini 2.5 Pro and achieving results comparable to Gemini 3 Pro.SINQopen sourceSINQ (Sinkhorn-Normalized Quantization) is a novel, fast, and high-quality quantization method designed to make any Large Language Model (LLM) smaller while preserving accuracy. It offers a plug-and-play, model-agnostic technique that delivers state-of-the-art performance for LLMs without sacrificing accuracy.