#inference services

37 stories · 6 in 7 days · 24 in 30 days
Stories, products and papers tagged "inference services", newest first.
Nov 13
Nov 11
Chinese content论文揭示前缀缓存的智能体服务存在亚稳态:一次突发可致吞吐腰斩70分钟

跑过 vLLM 智能体服务的人应该都见过那种莫名的性能崩塌,这篇论文把它解释清楚了:一次几小时的突发流量就能让命中率掉到 10%,吞吐腰斩还自己好不了。论文给了能用的预算器方案,不用调参,直接把 KV 池撑爆的可能性掐掉。

arXiv cs.LG@Wen Huang, Jack LiuSource
Chinese contentDatabricks 研究副总裁谈安全:不使用质量差的 neocloud GPU

这是 Databricks 内部人的第一视角:以前嫌公司安全流程太慢,看到别家推理服务商接连出安全事故后改观了。他说的具体一条是不用 neocloud 的低价 GPU。做企业服务、要评估推理供应商的团队值得看看这条思路。

Yuchen Jin@Yuchenj_UWSource
Nov 8
Nov 7
IndustryBuildersSelected04:39
OpenRouter Logs 790T Tokens in October, Up 25x Since January

The 25x token growth in ten months is a rough proxy for how fast real application traffic, especially agentic workloads, is scaling across the ecosystem. Developers choosing routing providers or pricing tiers, and analysts tracking which model labs retain usage, can mine the recap for demand signals. Compare it against your own token spend to benchmark adoption curves.

OpenRouter@OpenRouterAISource
Nov 5
IndustryUS vs ChinaSelected10:25
DeepSeek Tokens Split Across 20 Inference Providers

Open-weight models decouple the model from the provider, so serving capacity becomes a commodity market where dozens of hosts compete on price and latency. Teams deploying DeepSeek can shop across providers rather than depend on one vendor. The 17% self-hosted share shows even the model's creator holds only a minority of its own serving traffic.

AI Will@FinanceYF5Source
Nov 4
Nov 1
Oct 30
Oct 26
Oct 23
Oct 21
Oct 20
ProductsBuildersSelected13:07
PhysisServe Launches as Autoregressive Denoising Inference System for vLLM-Omni

This signals that diffusion-style and autoregressive denoising workloads are getting dedicated serving infrastructure rather than piggybacking on standard LLM stacks. Engineers running multimodal or video generation pipelines should watch vLLM-Omni, since the pipelining and multi-stream scheduling tricks could cut tail latency without a separate deployment. vLLM's direct endorsement also suggests upstream adoption of these scheduling techniques is likely.

Official accountvLLM@vllm_projectSource
Oct 15
Oct 8
Oct 4
ProductsBuildersMulti-sourceSelected04:18
Prime Intellect launches Prime Inference serving GLM-5.3 on vLLM

Serving agentic workloads reliably is a systems problem beyond raw token throughput, and this launch shows how topology and scheduler design affect tool-call reliability. Practitioners building agent infrastructure can study the approach or evaluate Prime Inference against self-hosted vLLM deployments. GLM-5.3, a Zhipu (Z.ai) model, being served at scale also signals growing demand for open Chinese models in Western inference stacks.

Event
Official accountvLLM@vllm_projectSource
Sep 30
Sep 29
Sep 22
Sep 19
Sep 9
Aug 26
Aug 13
Jul 28
Jul 17
May 19