ElasticMM – 4.2× Faster Multimodal LLM Serving (NeurIPS 2025 Oral)

Show HN: ElasticMM – 4.2× Faster Multimodal LLM Serving (NeurIPS 2025 Oral)

Based, ElasticMM – 4.2× Faster Multimodal LLM Serving (NeurIPS 2025 Oral) works in show hn launch.

Profile

ElasticMM is a newly released open-source serving system designed for modern multimodal large language models (MLLMs). The work was selected as an Oral presentation at NeurIPS 2025.Unlike existing serving stacks such as vLLM—which are primarily optimized for text-only workloads—ElasticMM introduces Elastic Multimodal Parallelism (EMP), a new execution paradigm that adapts parallelism across different inference stages and modalities.Key findings from the paper:Up to 4.2× reduction in TTFT3.2×–4.5× higher throughput under mixed multimodal workloadsModality-aware scheduling, elastic stage partitioning, unified prefix caching, and non-blocking encodingPaper (OpenReview PDF): https://openreview.n

As a business in show hn launch, ElasticMM – 4.2× Faster Multimodal LLM Serving (NeurIPS 2025 Oral) is listed here with the public details available about it.

The company is listed at the Show HN launch stage of development.

Website: openreview.net/pdf?id=Zd6VyjmN1SGitHub