Recent benchmark updates through mid-October 2026 have kept the third-place slot among Chinese AI labs highly contested, with Xiaomi’s MiMo-V2.6-Pro posting the highest scores on Artificial Analysis and related indices, followed closely by Alibaba’s Qwen3.8 Max and Z.ai’s GLM-5.3. Moonshot’s Kimi K3 remains competitive on long-context tasks and open-weight releases, while DeepSeek’s frequent updates and large funding round sustain its visibility despite trailing on some composites. Trader odds reflect this tight pack, as rapid model iterations from multiple labs—including StepFun and MiniMax—continue to shift relative positioning ahead of November resolution. Key swing factors include any new frontier-level releases or major benchmark refreshes before month-end.