Cherry-Pick Benchmarks

AI vendors often select specific benchmarks and outdated or closed models to highlight superior performance, leading to misleading comparisons and underlining the need for fair open-source benchmarks.