Iluvatar CoreX Tiangai 300 tops NVIDIA H200 in select AI inference tests, with long-context and MoE gains

Chinese AI chipmaker Iluvatar CoreX released benchmarks showing its next-generation Tiangai 300 GPU outperforming Nvidia's H200 in selected LLM inference, long-context processing, and multimodal generation workloads. The results highlight gains in long-context and mixture-of-experts (MoE) scenarios, positioning the chip as a competitive alternative in the AI inference market. The announcement reflects ongoing efforts by Chinese firms to challenge Nvidia's dominance in AI hardware.