DeepSeek DeepSpec: open-source speculative decoding
DeepSeek published DeepSpec, a full-stack MIT-licensed codebase for training and evaluating draft models for speculative decoding, bundling DSpark, DFlash, and Eagle3.
caveman: terse-output skill, 75k stars in 11 weeks
GitHub JuliusBrussee/caveman: a TypeScript skill for 30+ agent platforms that asks the agent to drop filler. 75k stars, MIT, 65% output-token reduction.
DiffusionGemma: 1,000+ tokens/sec open-weights text gen
Google DeepMind's June 10, 2026 release: a 26B/3.8B text-diffusion model denoising 256 tokens in parallel; ~4x faster than AR on a single H100, Apache 2.0, explicitly experimental.