vLLM Announces Day 0 Support for Kimi K3: Run 2.8T MoE on 8 B300 GPUs
vllm_project · x · 2026-07-30
The vLLM team announced efficient, production-scale Day 0 support for Moonshot AI's newly open-sourced Kimi K3 weights.
Architecture & Challenges
- Scale: Kimi K3 is a 2.8-trillion-parameter Mixture-of-Experts model with 16 active experts per token, a 1M-token context window, and native vision.
- Core Design: Built on Kimi Delta Attention (KDA) and Attention Residuals (AttnRes).
- Engineering: The team integrated KDA, MXFP4 MoE, KV cache management, prefill/decode disaggregation, speculative decoding, and long-context recipes into a runnable engine.
Deployment
The easiest deployment runs on 8 NVIDIA B300 or AMD MI355X GPUs. Users can also enable the open-sourced Inferact DSpark speculator via specific flags to boost inference speed.
Related event: Kimi K3 Gets Day-0 vLLM and AMD Support Across Clouds(14 posts)→
More from Infra
- Qualcomm goes agent-centric: Snapdragon 8 Elite Gen 6 and agent-native devices — jiqizhixin · 2026-09-23
- Unsloth Desktop Hotfix Adds Qwen-Image-2.1 Image Editing and Fixes GGUF Loading — danielhanchen · 2026-09-23
- Qwen 3.6 35B-A3B Q6 hits ~50 tok/s on a 128GB Strix Halo — what's the best local model now? — jankeydankey · 2026-09-23
- Together AI adds canary rollouts for zero-downtime model upgrades on dedicated inference — togethercompute · 2026-09-23
- Dedicated Hardware for Running AI Agents at Scale Arrives — cyrilzakka · 2026-09-23
- Ternary Bonsai 2 27B: 5.9GB weights retain ~95% of full-precision reasoning — cephaloform · 2026-09-23