Google's unparalleled compute and DeepMind's R&D drive SOTA. Gemini's scaling and specialized math fine-tuning show strong trajectory on benchmarks like MATH/GSM8K. Expect performance leadership. 90% YES — invalid if a competitor releases a new foundation model before April 30.
Xiaomi's AI strategy centers on IoT/mobile, not foundational coding LLMs. Zero competitive HumanEval/MBPP benchmark presence. Industry discourse firmly places them outside the top 5 contenders. 99% NO — invalid if a stealth Code Llama competitor is unveiled P2.
BOSS's dominant 2-0 potential dictates Even total rounds. Projecting 16-9, 16-11 map wins for a 52-round total. This aggressive sweep market signal is clear. 85% NO — invalid if the series extends beyond two maps.
