OpenBench provides standardized, reproducible benchmarking for LLMs across 30+ evaluation suites (and growing) spanning knowledge, math, reasoning, coding, science, reading comprehension, health, long ...

Domestic self-developed open-source modelsnow allow models to switch freely between fast and slow thinking! Huawei has ...