vLLM
vLLM appears in 39 Vector Wire stories since July, 11 of them about vLLM itself.
Coverage by Week
Since July 27 · busiest week Oct 5, 7 stories
Stories About vLLM
11- NVIDIA Dynamo Adds Session-Aware Routing for Agentic Inference Across vLLM, SGLangAlert · Oct 8
- vLLM cache_salt flaw lets one request crash EngineCore on LMCache-MP deploymentsBrief · Oct 6
- vLLM Patches Denial-of-Service Flaw in GLMGA Video SamplingAlert · Oct 6
- vLLM Discloses Unauthenticated DoS Flaw in Qwen Video BackendsAlert · Oct 6
- Helion Backend Outperforms CUTLASS, DeepGEMM in vLLM Inference on Hopper GPUsAlert · Oct 2
- OpenRLHF v0.11.1 Patches PPO NaN Bugs, Upgrades vLLM to 0.29.0Alert · Sep 10
More stories about vLLM
- vLLM Patches SSRF and Local File Read Bypassing Prior MediaConnector FixesBrief · Sep 8
- Red Hat Details FastMTP Heads to Speed vLLM Speculative DecodingAlert · Sep 8
- PyTorch Conference NA 2026 Dedicates Broad Track to vLLM Inference StackAlert · Aug 28
- vLLM v0.28.0 Ships Kimi-K3 and DeepSeek V4 OptimizationsAlert · Aug 26
Also Mentioned In
28- DeepSeek open-sources DeepInfer, the inference stack behind its production modelsOct 21
- Google-Meta TorchTPU opens native PyTorch path to TPUsOct 20
- PyTorch Foundation Adds Nebius, Nscale, Shopify as Platinum MembersOct 20
- AWS TorchNeuron Registers Trainium as Native PyTorch BackendOct 19
- Reflection ships Beam's 501B open weights under Apache 2.0, draws 40-plus launch partnersOct 18
More stories that mention vLLM
- Meta's Muse Spark Open bets 816B parameters on agentic edge over Chinese open-weight rivalsOct 17
- LiquidAI ships LFM2.5-24B-A2B, a 24B reasoning model that activates only 2B parametersOct 14
- Production failures become training data as enterprise AI loops widenOct 12
- Disaggregated Inference Moves from Design Pattern to Production DefaultOct 12
- Dynamo v1.5.1 Patches Router Overload, Multimodal Input BoundsOct 7
- Unsloth Adds Built-in Browser Panel, EmbeddingGemma 2, Audio Tools in v0.1.903 BetaOct 6
- PyTorch Working Group Ships Cross-Repo CI, Test Refactors for Hardware Backends in H1 2026Oct 5
- PyTorch Foundation Adds Introduction Track, Training Day to October ConferenceSep 24
- PyTorch Foundation Opens Call for Academic Projects Ahead of 2026 North America ConferenceSep 23
- Shopify Cuts AI Serving Costs 96% With Continual Learning LoopSep 22
- PyTorch Foundation Sets Tokyo Event for December 10Sep 18
- Quantization moves from checkpoint to serving engine as practitioners tune per workloadSep 10
- Alibaba Cloud, Cambricon, Ant Group Join PyTorch FoundationSep 8
- NVIDIA Dynamo Ships DeepSeek V4-Pro-0813 Recipes for GB200 and H200Sep 4
- PyTorch Conference NA 2026 Centers Agentic AI Across Two-Day ProgramSep 2
- Consumer GPUs now serve 27B models past 100 tok/s as software eats the inference moatSep 1
- NVIDIA Dynamo Adds Qwen 2.4T Serving on GB300, GB200Aug 27
- Unsloth Adds Local Inference for Qwen3.8-Flash-Next and GLM-5.3-FlashAug 27
- PyTorch Ecosystem Adds 10 Projects Spanning RL, Inference, Survival AnalysisAug 26
- Quantization and Hybrid Execution Push 100B+ Models onto Single Consumer GPUsAug 25
- AI-Dynamo Releases Dynamo v1.4.1 With OpenAI-Compatible Classify and Pooling EndpointsAug 22
- Inference Cost Engineering: Tensor Management, Storage Tiering, and RAG Routing ConvergeAug 19
- Moonshot Launches Kimi K3, 2.8T-Parameter Open Model Claiming Parity With Anthropic Opus 4.8Jul 17