Nick (@gpus-market) | Hashnode Toggle Sidebar Feed Pro Search Theme Sign in MoreDarkshift - a dark factory for softwareBug0 - The AI-native e2e QA regression testingThe foreword by Hashnode - official blog from the Hashnode teamPassmark - The open-source AI framework for regression testingHashnode gql skill - let your AI agent publish to your Hashnode blogHackathonsChangelogBrand@hashnode on XHashnode on LinkedInSupport - [email protected] of ConductTermsPrivacySitemap Search Hashnode Search posts, tags, users, and pages @gpus-market Nick @gpus-market·Joined June 2026 Building GPUs.market - One marketplace for dedicated GPU capacity. WebsiteShare About GPUs.market helps teams rent production-ready GPU infrastructure across a verified global capacity network. Choose the GPU, region, pricing model, and deployment shape that fits your workload, then launch dedicated hardware with SSH access and full control over your environment. Available for Nothing here yet. Nick's blogs GPUs.market Engineering blogblog.gpus.market4 posts About GPUs.market helps teams rent production-ready GPU infrastructure across a verified global capacity network. Choose the GPU, region, pricing model, and deployment shape that fits your workload, then launch dedicated hardware with SSH access and full control over your environment. Available for Nothing here yet. Nick's blogs GPUs.market Engineering blogblog.gpus.market4 posts ArticlesComments Recently published NNickinblog.gpus.market·19h ago · 9 min read Deploying Qwen3.8-2.4T-A95B with vLLM: Verified GPU Pods, Quants, and Serving Recipes Qwen3.8-2.4T-A95B is a 2.4-trillion-parameter Mixture-of-Experts model with roughly 95B parameters active for each token. If you're planning to self-host it, the first thing to know is that this is a 10 NNickinblog.gpus.market·20h ago · 10 min read Deploying Kimi K3 with vLLM: Verified GPU Pods, Quants, and Serving Recipes Kimi K3 is Moonshot AI's 2.8-trillion-parameter Mixture-of-Experts model for coding, reasoning, agents, and multimodal work. It activates 16 of 896 routed experts for each token and supports a native 10 NNickinblog.gpus.market·21h ago · 10 min read Running GLM-5.2 in Production with vLLM: Quants, GPU Pods, and Verified Serving Recipes GLM-5.2 is Z.ai’s 743B-parameter Mixture-of-Experts model for coding, reasoning, and long-running agent workflows. Only about 39B parameters are active for each token, and the model supports a native 10 NNickinblog.gpus.market·22h ago · 12 min read Deploying DeepSeek V4 Flash 0731 with vLLM: Verified GPU Pods, Quants, and Serving Recipes DeepSeek V4 Flash 0731 is the current official Flash release: deepseek-ai/DeepSeek-V4-Flash-0731 It keeps the same underlying architecture as the earlier V4 Flash preview, but DeepSeek re-trained the 10