Models & Labs
vLLM v0.29.0 Release
vLLM Releaseshigh confidence
Why it matters
- →Model Runner V2 improves performance and memory efficiency for all models.
- →New models and optimizations reduce latency and enhance processing capabilities.
- →Transitioning from MRV1 to MRV2 will streamline future updates and support.
vLLM has released version 0.29.0, featuring Model Runner V2 as the default for all models, which enhances performance and memory management. This update includes new models such as Hy4-preview and Tencent's MoE, along with improvements in speculative decoding and RL weight synchronization. Performance optimizations for Kimi-K3 and DeepSeek V4 are also part of this release, promising reduced latency. While Model Runner V1 is being deprecated, vLLM plans to address remaining feature gaps in MRV2 shortly, enhancing its utility for developers.
Read originalThe story around this
Earlier coverage that leads up to this article, and what followed. Lines connect each piece to the closest one after it, converging here.
How we got here
- 1llama.cpp b9330 release improves model performance
llama.cpp Releases · May 27, 2026 · Related
- 2vLLM v0.22.0 Release Enhances Model Performance
vLLM Releases · May 31, 2026 · Same story
- 3Run vLLM Server on HF Jobs with One Command
Hugging Face Blog · June 26, 2026 · Related
- 4vLLM v0.24.0 Release Enhances Model Support
vLLM Releases · June 30, 2026 · Same story
- 5vLLM Boosts Transformers Backend for Native-Speed Inference
Hugging Face Blog · July 8, 2026 · Related
What happened next
- 6llama.cpp b11222 release with ROCm 10 and CUDA 13
llama.cpp Releases · September 28, 2026 · Related
Follow this story
Open the full story →vLLM v0.29.0 Release
2 developments
- Sep 11 · vLLM Releases
vLLM v0.29.0 Release (This article) - Sep 16 · vLLM Releases
vLLM v0.29.0rc6 fixes hybrid model caching↳ vLLM v0.29.0rc6 fixes hybrid model caching by enabling dense prefix cache by default
More in Models & Labs
Models & Labsmodels
llama.cpp b12029 adds CUDA 13 and Snapdragon support
This release quietly cements llama.cpp as the universal inference runtime by finally supporting NVIDIA's latest CUDA 13 stack alongside AMD's ROCm 10.0. The inclusion of native builds for Linux arm64 with Snapdragon chips, covering CPU, Adreno GPU, and Hexagon NPU, signals a serious push into mobile edge deployment. While Apple Silicon KleidiAI is disabled in this specific build, the broader hardware coverage means developers no longer need to compile from source for most modern accelerators. It’s less about new features and more about removing friction for anyone running local models on bleeding-edge or diverse hardware.
llama.cpp Releases
© The Verge AIModels & Labsother
Apple Intelligence arrives in China via iOS 27.3
Apple is finally unlocking its AI suite for Chinese users with the iOS 27.3 beta, marking a significant regulatory milestone after months of approval delays. The rollout relies on Alibaba’s Qwen model rather than OpenAI, integrating features like notification summaries and visual intelligence directly into the OS. By sidestepping the new Siri in favor of established text and image tools, Apple is prioritizing immediate utility over the full generative experience. This move effectively brings Apple’s AI capabilities to a massive new market, albeit with localized constraints that distinguish it from the global release.
The Verge AI
© Matt WolfeModels & Labsmodels
Google Releases Gemini 4 Carbon for Cost-Efficient Inference
Google launched Gemini 4 Carbon, a new model variant aimed at providing high performance at a lower cost for inference tasks.
Matt Wolfe