🧰 TileRT

The items that name it 3

• [AINews] Opus 5.5 is good at explainer videos
  article2026-09-29

  Used alongside vLLM for disaggregated prefill/decode serving of GLM-5.3

• How GLM5.3 Sparse Attention Affects HBM Memory Usage
  article2026-09-28

  Low-latency LLM inference engine that compiles decoding into a single persistent kernel, favouring per-user token speed over batched throughput.

• [AINews] The Future of Latent Space
  article2026-09-25

  used alongside vLLM for disaggregated prefill/decode serving of GLM-5.3