2026​

• February 4 - llm-d 0.5: Sustaining Performance at Scale
• February 10 - Native KV Cache Offloading to Any Filesystem with llm-d
• March 13 - Predicted-Latency Based Scheduling for LLMs
• April 21 - Production-Grade LLM Inference at Scale with KServe, llm-d, and vLLM
• May 12 - llm-d v0.7: From Feature Introduction to Production Hardening
• June 5 - BLIS: Evolving llm-d at Simulation Speed
• June 9 - Heterogeneous inference serving across three GPU vendors with llm-d
• June 13 - Serving Hybrid Models at Scale in llm-d
• June 23 - Networking for Distributed Inference in llm-d

2025​

• May 20 - llm-d Press Release
• May 20 - Announcing the llm-d community!
• June 3 - llm-d Week 1 Project News Round-Up
• June 25 - llm-d Community Update - June 2025
• July 29 - llm-d 0.2: Our first well-lit paths (mind the tree roots!)
• September 3 - Intelligent Inference Scheduling with llm-d
• September 24 - KV-Cache Wins You Can See: From Prefix Caching in vLLM to Distributed Scheduling with llm-d
• October 10 - llm-d 0.3: Wider Well-Lit Paths for Scalable Inference
• December 2 - llm-d 0.4: Achieve SOTA Performance Across Accelerators