# Researchers propose tracequery to cut diffusion LLM inference costs

Digest AI · Research · published 2026-10-27T04:00:00Z

Canonical: https://digestai.news/story/researchers-propose-tracequery-to-cut-diffusion-llm-inference-costs

## Summary

A new method called TRACEQUERY aims to reduce the computational cost of diffusion-based large language models (dLLMs) by focusing recomputation on relevant parts of the generation trajectory. The paper, posted on arXiv, argues that existing cache-based approaches waste resources by recomputing tokens on a fixed schedule or based on heuristics. Instead, TRACEQUERY uses full-denoising trajectories to predict which positions need recomputation, minimizing fresh calculations while maintaining accuracy.

The authors claim TRACEQUERY outperforms other cache-based methods in balancing quality and computation efficiency. It works across tasks, model variants, and families, according to the paper. The code is available on GitHub under the repository [tracequery](https://github.com/fish1928/tracequery.git).

## Key points

- TRACEQUERY predicts trajectory-relevant positions to reduce token recomputation in dLLMs
- Method claims stronger quality-computation trade-off than existing cache-based approaches
- Code released on GitHub for further research and implementation

## Why it matters

If validated, TRACEQUERY could lower inference costs for diffusion models, making them more practical for real-time applications.

## Sources

1. [TraceQuery: learning what to recompute from full-denoising trajectory in diffusion language models](https://arxiv.org/abs/2610.33008) (arXiv cs.CL, 2026-10-27, primary source)

## Cite

Digest AI, "Researchers propose tracequery to cut diffusion LLM inference costs", 27 October 2026, https://digestai.news/story/researchers-propose-tracequery-to-cut-diffusion-llm-inference-costs

---

Written by Digest AI's editorial model from the linked sources; the sources are the record. Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse
JSON: https://digestai.news/story/researchers-propose-tracequery-to-cut-diffusion-llm-inference-costs.json