Kimi K3 Launches with Day 0 Support on AMD Instinct via vLLM
vllm_project · x · 2026-07-30
vLLM officially announced that the Kimi K3 model supports AMD Instinct hardware from day zero. Thanks to vLLM's adaptation to ROCm, developers can efficiently deploy and serve the full 2.8 trillion parameter model on AMD hardware right from the start. Broader performance tuning for AMD Instinct is also actively in progress.
Related event: vLLM Day-0 Support for Kimi K3: Run 2.8T MoE on 8 B300 GPUs(11 posts)→
More from Infra
- Goldman Sachs Predicts 70x Jump in Monthly AI Token Processing by 2030 — Beth_Kindig · 2026-07-30
- Cold Start Benchmark: 244 GiB Model Loads in 154 Seconds — QuixiAI · 2026-07-30
- Unified FP8 in Training and Rollout Speeds Up RL by 16% — joecole · 2026-07-30
- ThunderAgent Engine: 2× Throughput, Near-Linear Multi-Node Scaling for Agent Workflows — togethercompute · 2026-07-30
- ThunderAgent (ICML 2026 Spotlight): Overcomes KV Cache Thrashing in Agentic Inference — togethercompute · 2026-07-30
- QuixiCore-ROCm Open Source: High-Performance Kernel Library Tuned for AMD MI300x — QuixiAI · 2026-07-30