What field is speculative decoding part of?

Speculative decoding is part of Efficient inference and compression.

Is speculative decoding ready for products?

Its papers score 2.5 of 5 for readiness for products, read from what the papers themselves describe.

Which companies work on speculative decoding?

Researchers at Tencent publish on it.

Speculative decoding

Part of Efficient inference and compression

Speculative decoding is gaining ground in AI research: 0.3% to 0.4% of new AI papers, +0.0 points. It ranks #57 of 214 by gain in share. As of Oct 10, 2026.

How we measure

Change in share: +0.0 pts
Share, last 3 months: 0.4%
Papers, last 3 months: 125
Ready for products: 2.5 / 5

More in Efficient inference and compression

Kv cache reuse0.1% → 0.2%+0.1 pts Kv cache compression0.2% → 0.3%+0.1 pts Load balancing0.1% → 0.1%+0.1 pts Visual token compression0.1% → 0.1%+0.1 pts Neural network quantization0.0% → 0.1%+0.0 pts Diffusion model acceleration0.2% → 0.2%+0.0 pts Structured pruning0.1% → 0.1%+0.0 pts Mixed precision quantization0.1% → 0.1%+0.0 pts Visual token pruning0.2% → 0.2%+0.0 pts Adaptive computation0.2% → 0.2%+0.0 pts

Companies working on it

1Tencent1 papers

Speculative decoding: quick answers

Yes. Its share of new AI papers went from 0.3% to 0.4%, with 125 papers in the last 3 months.