Inception Labs released Mercury 2, the first reasoning diffusion LLM (dLLM). The standout: 1009 tokens/sec on a single H100, a 10× speedup over autoregressive (AR) reasoning models, with comparable quality on math and code reasoning.
The "reasoning diffusion" innovation: Mercury 2 is a diffusion language model (generates all tokens in parallel via iterative denoising) that has been specifically trained for "reasoning" — the model generates a "thought" before the final answer. The combination of diffusion's parallel generation and reasoning's "think before answering" gives a 10× speedup over AR reasoning models.
The technical details: Mercury 2 uses a "thought-conditioned" diffusion process. The model first generates a "thought" representation in continuous space (no decoding cost), then uses this thought to guide the diffusion process for the final answer. The result is reasoning-quality output at diffusion speed.
The benchmark: on the MATH benchmark, Mercury 2 scores 87.4, on par with GPT-5.6 (87.4) and Claude Opus 4.7 (88.1). The 1009 tokens/sec throughput is 10× faster than GPT-5.6's ~100 tokens/sec on the same hardware.
The "real-time Agent cost curve" highlight: 1009 tokens/sec changes the economics of real-time Agents. At 100 tokens/sec, a 10-second response requires 1000 tokens — affordable but limited. At 1009 tokens/sec, a 10-second response can be 10,000 tokens — enabling much richer responses, longer context handling, and more complex multi-step reasoning.
The bigger takeaway: "reasoning diffusion LLMs" are a real paradigm shift. The "AR is the only way to reason" assumption is being broken, and Mercury 2's 10× speedup is a clear signal. For the industry, this means "real-time reasoning Agents" (code generation, multi-step planning, complex QA) are now economically viable, and the vendors that adopt reasoning dLLMs first will have a significant cost advantage.