vLLM v0.24.0 Release: Adds MiniMax-M3 Support, BF16/FP8 Indexer

Core summary

One sentence to understand this update

vLLM v0.24.0 has been released, featuring support for the new MiniMax-M3 model and a rapid follow-up for BF16/FP8 indexer via MSA.

Impact & opportunity

What this could mean

Developers using vLLM can now leverage the latest MiniMax-M3 model and benefit from improved performance with BF16/FP8 indexing for more efficient large language model serving.