Updates

Two new multimodal models are now available on Dell Enterprise Hub. Qwen3.8-Flash-Next-FP8, a MoE model with 125B parameters and just 6B active parameters, is the preview version of Qwen4, pairing efficient sparse attention with FP8 weights for long-context workloads. Meanwhile, GLM-5.3-Flash, a 320B MoE model with 18B active parameters, is built for coding and agentic tasks. Both frontier open models achieve great performance and you already can deploy them on-premise on Dell platforms.

Deploy Qwen3.8 Flash Next FP8 → · Deploy GLM 5.3 Flash →

dell-ai v1.0 is a Python SDK and CLI for Dell Enterprise Hub. It lets you discover validated models, generate Docker, Kubernetes, or Helm configs, and deploy locally with automatic GPU allocation. It even comes with an agent skill so tools like Cursor and Claude Code can run the same workflow from natural language. Go take a look at the release announcement and start using it today!

Read the announcement → · View on GitHub →

Qwen3.8-27B is now available on Dell Enterprise Hub. A compact 27B dense vision-language model with native image and video understanding, flexible thinking control, and stronger coding, research, and long-horizon agentic performance. Deploy it with vLLM 0.27.1 on Dell Pro Max with GB10, 4×L40S, or 1×, 2×, and 4× H100 and H200 systems.

Deploy Qwen3.8 27B →

DeepSeek V4 Pro 0813 is now available on Dell Enterprise Hub. This updated 1.6T-parameter Mixture-of-Experts model with 49B activated parameters delivers advanced long-context reasoning and agentic capabilities. Deploy it on-prem with vLLM 0.27.1 on H200 and B300 systems.

Deploy DeepSeek V4 Pro 0813 →

The newest NVIDIA model, Nemotron 3.5 Lightning 30B A3B, is now available in NVFP4 and BF16 formats on Dell Enterprise Hub. Deploy this model with great agentic capabilities and speculative decoding powered by DSpark and DFlash on a great range of Dell platforms.

Deploy NVFP4 → · Deploy BF16 →