HOT RADAR
llmHEAT 12.2°

AI CURATED HOT STORY · 11/20/2026

NVIDIA Releases Sigma Diffusion Language Models with 3.46x Faster Decoding

1 reports archived1 independent sourcesupdated 11/20/2026, 22:50:42
Synthesis & Latest Updates
1 Sources Cross-Validated

NVIDIA released Nemotron-Labs-Sigma, 14 continuous diffusion language model checkpoints at 3B and 8B that move diffusion out of discrete token space into a learned 16-D embedding space, adding and removing Gaussian noise block by block. Sigma-8B averages 71.16 on NVIDIA's 6-task base suite versus 70.22 for the discrete-diffusion NLD-8B, and self-speculative decoding runs 3.46x faster on average (up to 4.20x) on a single GB200. The trade-off is clear: instruct average of 59.14 trails NLD-8B's 59.92, and AIME25 sits at just 17.00.

LATEST/NVIDIA released Nemotron-Labs-Sigma, 14 continuous diffusion language model checkpoints at 3B and 8B that move diffusion out of discrete token space into a learned 16-D embedding space, adding and removing Gaussian noise block by block. Sigma-8B averages 71.16 on NVIDIA's 6-task base suite versus 70.22 for the discrete-diffusion NLD-8B, and self-speculative decoding runs 3.46x faster on average (up to 4.20x) on a single GB200. The trade-off is clear: instruct average of 59.14 trails NLD-8B's 59.92, and AIME25 sits at just 17.00.

Hourly heat curve

14 fully observed hours
Now
12.2
Peak
16.711-21
24h change
-27%

The line compares only sources observed throughout; gap hours are interpolated to keep the trend continuous.

Coverage timeline

Total 1 reports · Latest first
  1. Marktechpost (@Marktechpost)T2·78 pts
    • Moves diffusion from discrete token space to a learned 16-D embedding space, where smooth trajectories make few-step distillation and steering practical
    • Sigma-8B scores 71.16 on the base suite vs 70.22 for NLD-8B, and 65.12 vs 62.80 on HumanEval
    • Self-speculative decoding is 3.46x faster on average on one GB200, but instruct following and AIME25 reasoning lag behind